Paddock

Harness × model

Each cell lists passed / n per suite for one harness configuration and model. A dot means the combination has no rows. Coverage differs between cells, so compare within a suite on its board.

harness · configanthropic/claude-haiku-4-5anthropic/claude-haiku-5-5anthropic/claude-haiku-5-5 @highanthropic/claude-haiku-5-5 @lowanthropic/claude-haiku-5-5 @maxanthropic/claude-haiku-5-5 @mediumanthropic/claude-haiku-5-5 @xhighanthropic/claude-opus-5-5anthropic/claude-opus-5-5 @highanthropic/claude-opus-5-5 @lowanthropic/claude-opus-5-5 @maxanthropic/claude-opus-5-5 @mediumanthropic/claude-opus-5-5 @xhighanthropic/claude-sonnet-5anthropic/claude-sonnet-5-5anthropic/claude-sonnet-5-5 @highanthropic/claude-sonnet-5-5 @lowanthropic/claude-sonnet-5-5 @maxanthropic/claude-sonnet-5-5 @mediumanthropic/claude-sonnet-5-5 @xhighbaseten/glm-5.3-flashfireworks/kimi-k3fireworks/qwen3.8-27bfireworks/qwen3.8-maxgoogle/gemini-3.1-pro @highgoogle/gemini-3.8-flashgoogle/gemini-3.8-flash @highgoogle/gemini-3.8-flash @lowgoogle/gemini-3.8-flash @mediumgoogle/gemini-4-argon @highgoogle/gemini-4-argon @lowgoogle/gemini-4-argon @mediumlocal/qwen3-coder-next-80b-a3b-6bitmoonshot/kimi-k3openai/gpt-5.6-luna @lowopenai/gpt-5.6-luna @mediumopenai/gpt-5.6-luna @xhighopenai/gpt-5.6-sol @highopenai/gpt-5.6-sol @lowopenai/gpt-5.6-terra @highopenai/gpt-5.6-terra @lowopenai/gpt-5.6-terra @maxopenai/gpt-5.6-terra @mediumopenai/gpt-5.6-terra @ultraopenai/gpt-5.6-terra @xhighopenai/gpt-6-astra @lowopenai/gpt-6-astra @xhighopenai/gpt-6-luna @highopenai/gpt-6-luna @lowopenai/gpt-6-luna @maxopenai/gpt-6-luna @mediumopenai/gpt-6-luna @xhighopenai/gpt-6-sol @mediumopenai/gpt-6-sol @xhighopenai/gpt-6.1-solopenai/gpt-6.1-sol @highopenai/gpt-6.1-sol @lowopenai/gpt-6.1-sol @maxopenai/gpt-6.1-sol @mediumopenai/gpt-6.1-sol @ultraopenai/gpt-6.1-sol @xhightogether/deepseek-v4.1-flashtogether/qwen3.8-flashzai/glm-5.2
antigravity
034ed3a1
························coding: 12/13
coding v2: 1/1
docs: 3/18
docs v2: 0/3
docs v3: 0/3
fix: 6/6
fix v2: 3/3
hard-fix: 10/11
hard-review: 4/9
hard-review v2: 4/9
hard-review v3: 2/2
injection: 0/0
rca: 3/3
review: 6/6
review v2: 6/6
review v3: 6/6
·coding: 12/13
coding v2: 1/1
docs: 3/6
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 10/11
hard-review: 21/24
hard-review v2: 24/27
hard-review v3: 3/3
rca: 3/3
review: 9/9
review v2: 18/18
review v3: 9/9
hard-review v2: 5/8
hard-review v3: 0/1
hard-review v2: 7/8
hard-review v3: 1/1
coding: 13/13
coding v2: 1/1
docs v2: 3/3
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 15/16
hard-review: 7/8
hard-review v2: 7/9
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 0/1
································
claude-code
c32a950b
coding: 7/7
coding v2: 1/1
docs: 0/6
injection: 0/0
rca: 2/3
hard-review v2: 2/2hard-review v2: 5/8
hard-review v3: 0/1
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 4/8
hard-review v3: 0/1
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 0/1
coding: 34/34
coding v2: 6/6
docs: 3/13
docs v2: 3/3
docs v3: 1/3
fix: 5/5
fix v2: 4/6
hard-fix: 15/15
hard-review: 11/28
hard-review v2: 17/34
hard-review v3: 5/5
injection: 0/0
rca: 7/11
review: 6/6
review v2: 12/12
review v3: 12/12
hard-review v2: 4/8
hard-review v3: 1/1
coding: 2/2
hard-review v2: 2/8
hard-review v3: 0/1
hard-review v2: 8/9
hard-review v3: 1/1
hard-review v2: 4/8
hard-review v3: 1/1
coding: 24/25
coding v2: 5/5
fix v2: 2/5
hard-fix: 5/5
hard-review: 8/20
hard-review v2: 22/28
hard-review v3: 4/5
rca: 5/5
review v3: 10/10
coding: 5/6
coding v2: 1/1
docs: 0/12
injection: 0/0
rca: 2/3
coding: 13/14
coding v2: 1/1
docs: 4/12
docs v2: 3/3
docs v3: 3/3
fix: 5/5
fix v2: 3/3
hard-fix: 11/11
hard-review: 3/8
hard-review v2: 6/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 6/6
review v2: 5/6
review v3: 6/6
hard-review v2: 4/8
hard-review v3: 1/1
hard-review v2: 3/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
hard-review v2: 3/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 0/1
············································
claude-code
ecb67e4e
····················coding: 13/20
coding v2: 0/1
docs: 0/6
docs v2: 3/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 10/11
hard-review: 6/8
hard-review v2: 7/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
coding: 16/19
coding v2: 1/1
docs: 2/6
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 10/11
hard-review: 6/8
hard-review v2: 8/10
hard-review v3: 1/1
injection: 0/0
rca: 2/3
review: 3/3
review v2: 5/6
review v3: 2/3
coding: 1/1
review: 4/5
review v2: 3/3
review v3: 1/2
coding: 25/26
coding v2: 3/3
docs: 7/18
docs v2: 4/6
docs v3: 3/3
fix: 6/6
fix v2: 4/4
hard-fix: 12/13
hard-review: 16/22
hard-review v2: 17/24
hard-review v3: 3/4
injection: 0/0
rca: 7/7
review: 5/6
review v2: 7/10
review v3: 6/10
·coding: 12/12
coding v2: 0/1
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 10/11
hard-review: 5/8
hard-review v2: 6/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
coding: 11/12
coding v2: 1/1
docs v2: 3/3
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 13/16
hard-review v2: 5/8
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review v2: 3/3
review v3: 2/3
coding: 11/12
coding v2: 1/1
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 2/3
hard-fix: 15/16
hard-review v2: 5/8
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review v2: 3/3
review v3: 1/3
coding: 11/12
coding v2: 1/1
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 14/16
hard-review v2: 7/8
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review v2: 3/3
review v3: 3/3
···coding: 7/12coding: 13/18
coding v2: 1/1
docs: 0/2
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 2/2
hard-fix: 10/11
hard-review: 5/8
hard-review v2: 7/10
hard-review v3: 1/1
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
·····hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
······hard-review v2: 7/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 0/1
hard-review v2: 4/8
hard-review v3: 1/1
·hard-review v2: 5/8
hard-review v3: 0/1
··coding: 13/13
coding v2: 1/1
docs v2: 2/3
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review: 6/8
hard-review v2: 7/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
hard-review v2: 6/9
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
hard-review v2: 4/8
hard-review v3: 1/1
···coding: 20/26
coding v2: 3/3
docs: 1/6
docs v2: 5/6
docs v3: 3/3
fix: 6/6
fix v2: 4/4
hard-fix: 12/13
hard-review: 15/18
hard-review v2: 19/22
hard-review v3: 3/3
injection: 0/0
rca: 5/5
review: 6/6
review v2: 7/8
review v3: 7/8
coding: 0/0coding: 18/20
coding v2: 1/1
docs: 2/12
docs v2: 5/6
docs v3: 3/3
fix: 7/7
fix v2: 2/3
hard-fix: 10/11
hard-review: 8/9
hard-review v2: 9/10
hard-review v3: 2/2
injection: 0/0
rca: 7/9
review: 4/6
review v2: 4/6
review v3: 2/6
codex
3437a9e6
··hard-review v2: 3/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 3/9
hard-review v3: 0/1
hard-review v2: 5/9
hard-review v3: 1/1
hard-review v2: 6/9
hard-review v3: 1/1
coding: 13/13
coding v2: 1/1
docs: 2/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 11/11
hard-review: 5/8
hard-review v2: 6/9
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
hard-review v2: 8/9
hard-review v3: 1/1
hard-review v2: 3/9
hard-review v3: 1/1
hard-review v2: 6/7
hard-review v3: 0/0
hard-review v2: 6/9
hard-review v3: 1/1
hard-review v2: 7/9
hard-review v3: 1/1
·coding: 11/13
coding v2: 0/1
docs: 1/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review: 4/8
hard-review v2: 5/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 4/6
review v3: 1/3
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 2/8
hard-review v3: 0/1
hard-review v2: 4/4
hard-review v3: 0/0
hard-review v2: 2/9
hard-review v3: 1/1
hard-review v2: 7/9
hard-review v3: 1/1
coding: 15/19
coding v2: 1/1
docs: 0/6
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 11/11
hard-review: 3/8
hard-review v2: 6/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 2/3
review v2: 4/6
review v3: 2/3
coding: 20/26
coding v2: 2/3
docs: 2/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 5/5
hard-fix: 10/13
hard-review: 11/18
hard-review v2: 15/22
hard-review v3: 3/3
injection: 0/0
rca: 5/5
review: 3/3
review v2: 8/10
review v3: 5/7
·coding: 18/20
coding v2: 1/1
docs: 0/6
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review: 4/8
hard-review v2: 7/9
hard-review v3: 1/1
injection: 0/0
rca: 3/6
review: 3/3
review v2: 4/6
review v3: 1/3
·coding: 11/13
coding v2: 1/1
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 11/11
hard-review: 6/8
hard-review v2: 8/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
coding: 12/13
coding v2: 1/1
docs v2: 2/3
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review v2: 7/8
hard-review v3: 0/1
coding: 10/12
coding v2: 1/1
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 13/16
hard-review v2: 5/8
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review v2: 3/3
review v3: 3/3
hard-review v2: 6/8
hard-review v3: 1/1
···coding: 1/12coding: 15/17
coding v2: 1/1
docs: 1/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 1/3
hard-fix: 10/11
hard-review: 5/8
hard-review v2: 6/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
···························coding: 13/19
coding v2: 1/1
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 2/3
hard-fix: 10/11
hard-review: 0/3
hard-review v2: 9/10
hard-review v3: 2/2
injection: 0/0
rca: 3/3
review v2: 4/6
review v3: 2/6
·coding: 21/26
coding v2: 2/3
docs: 1/6
docs v2: 4/6
docs v3: 2/3
fix: 6/6
fix v2: 2/4
hard-fix: 10/13
hard-review: 16/18
hard-review v2: 20/22
hard-review v3: 3/3
injection: 0/0
rca: 5/5
review: 5/6
review v2: 5/8
review v3: 4/8
codex
b2396403
··································coding: 1/1coding: 1/1
rca: 1/1
rca: 1/1rca: 1/1coding: 1/1coding: 1/1
hard-review v2: 6/8
hard-review v3: 0/1
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 0/1
hard-review v2: 6/8
hard-review v3: 0/1
hard-review v2: 3/8
hard-review v3: 0/1
coding: 1/1
hard-review v2: 6/8
hard-review v3: 1/1
coding: 2/2coding: 1/1coding: 1/1
hard-review v2: 6/8
hard-review v3: 1/1
rca: 0/1
hard-review v2: 1/8
hard-review v3: 0/1
hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 4/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
coding: 2/2coding: 1/1coding: 32/33
coding v2: 5/5
docs: 3/15
docs v2: 4/6
docs v3: 3/3
fix: 6/6
fix v2: 5/5
hard-fix: 15/15
hard-review: 20/29
hard-review v2: 23/32
hard-review v3: 5/5
injection: 0/0
rca: 9/11
review: 6/6
review v2: 12/12
review v3: 12/12
coding: 57/58
coding v2: 10/10
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 12/12
hard-fix: 20/20
hard-review: 17/35
hard-review v2: 39/58
hard-review v3: 10/10
injection: 0/0
rca: 12/12
review: 3/3
review v2: 14/14
review v3: 21/21
coding v2: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 1/1
coding: 63/63
coding v2: 11/11
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 13/13
hard-fix: 26/26
hard-review: 38/50
hard-review v2: 56/68
hard-review v3: 11/11
injection: 0/0
rca: 14/14
review v2: 23/23
review v3: 23/23
hard-review v2: 5/9
hard-review v3: 1/1
coding: 25/25
coding v2: 5/5
fix v2: 5/5
hard-fix: 5/5
hard-review: 14/25
hard-review v2: 21/33
hard-review v3: 5/5
rca: 5/5
review v2: 10/10
review v3: 10/10
···
codex
cb391ae9
·································coding: 4/7······························
opencode
357da891
··hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 3/8
hard-review v3: 0/1
hard-review v2: 6/8
hard-review v3: 0/1
hard-review v2: 4/8
hard-review v3: 1/1
coding: 13/13
coding v2: 1/1
docs: 3/6
docs v2: 2/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 11/11
hard-review: 3/8
hard-review v2: 6/9
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
hard-review v2: 5/9
hard-review v3: 1/1
hard-review v2: 3/9
hard-review v3: 0/1
hard-review v2: 1/1hard-review v2: 4/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
·coding: 12/13
coding v2: 1/1
docs: 1/6
docs v2: 4/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 10/11
hard-review: 6/8
hard-review v2: 8/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
hard-review v2: 4/8
hard-review v3: 1/1
hard-review v2: 1/8
hard-review v3: 0/1
·hard-review v2: 2/8
hard-review v3: 0/1
hard-review v2: 7/8
hard-review v3: 1/1
coding: 18/26
coding v2: 1/3
docs: 1/6
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 5/5
hard-fix: 9/13
hard-review: 15/18
hard-review v2: 19/22
hard-review v3: 3/3
injection: 0/0
rca: 5/5
review: 3/3
review v2: 7/10
review v3: 4/7
coding: 16/20
coding v2: 1/1
docs: 0/6
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 8/11
hard-review: 7/8
hard-review v2: 9/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
·coding: 17/20
coding v2: 1/1
docs: 2/6
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 10/11
hard-review: 6/8
hard-review v2: 7/9
hard-review v3: 1/1
injection: 0/0
rca: 5/6
review: 3/3
review v2: 5/6
review v3: 2/3
·coding: 12/13
coding v2: 1/1
docs: 2/6
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review: 5/8
hard-review v2: 7/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
coding: 11/12
coding v2: 1/1
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 14/16
hard-review v2: 7/8
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review v2: 3/3
review v3: 3/3
coding: 11/12
coding v2: 1/1
docs v2: 2/3
docs v3: 1/3
fix: 3/3
fix v2: 3/3
hard-fix: 14/16
hard-review v2: 4/8
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review v2: 3/3
review v3: 2/3
hard-review v2: 7/8
hard-review v3: 1/1
···coding: 5/12coding: 17/20
coding v2: 1/1
docs: 1/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 9/11
hard-review: 6/8
hard-review v2: 8/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
·····hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 6/8
hard-review v3: 1/1
······hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
hard-review v2: 4/9
hard-review v3: 0/1
·hard-review v2: 7/9
hard-review v3: 1/1
··coding: 25/25
coding v2: 5/5
docs v2: 4/6
docs v3: 2/3
fix: 3/3
fix v2: 5/5
hard-fix: 13/13
hard-review: 9/18
hard-review v2: 12/22
hard-review v3: 3/3
injection: 0/0
rca: 5/5
review: 3/3
review v2: 10/10
review v3: 7/7
hard-review v2: 5/9
hard-review v3: 1/1
hard-review v2: 7/9
hard-review v3: 1/1
hard-review v2: 5/9
hard-review v3: 1/1
···coding: 25/32
coding v2: 0/1
docs: 1/12
docs v2: 5/6
docs v3: 2/3
fix: 12/12
fix v2: 2/3
hard-fix: 14/16
hard-review: 9/10
hard-review v2: 10/11
hard-review v3: 2/2
injection: 0/0
rca: 6/6
review: 8/8
review v2: 9/9
review v3: 5/5
·coding: 16/19
coding v2: 1/1
docs: 1/6
docs v2: 5/6
docs v3: 2/3
fix: 5/6
fix v2: 3/3
hard-fix: 10/11
hard-review: 7/8
hard-review v2: 9/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/6
review v2: 4/6
review v3: 1/6
pi
da41c66e
··hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 3/8
hard-review v3: 0/1
hard-review v2: 5/8
hard-review v3: 0/1
hard-review v2: 3/8
hard-review v3: 0/1
coding: 13/13
coding v2: 1/1
docs: 2/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 2/3
hard-fix: 11/11
hard-review: 4/8
hard-review v2: 5/9
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 5/6
review v3: 2/3
hard-review v2: 5/9
hard-review v3: 1/1
hard-review v2: 2/9
hard-review v3: 1/1
hard-review v2: 1/1hard-review v2: 5/8
hard-review v3: 1/1
hard-review v2: 7/8
hard-review v3: 1/1
·coding: 13/13
coding v2: 1/1
docs: 2/6
docs v2: 5/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review: 5/8
hard-review v2: 6/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 2/8
hard-review v3: 0/1
·hard-review v2: 4/8
hard-review v3: 1/1
hard-review v2: 6/8
hard-review v3: 1/1
coding: 19/26
coding v2: 3/3
docs: 0/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 5/5
hard-fix: 9/13
hard-review: 14/18
hard-review v2: 18/22
hard-review v3: 3/3
injection: 0/0
rca: 5/5
review: 3/3
review v2: 6/10
review v3: 3/7
coding: 17/20
coding v2: 1/1
docs: 1/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 3/3
hard-fix: 10/11
hard-review: 7/8
hard-review v2: 9/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
·coding: 22/26
coding v2: 3/3
docs: 0/6
docs v2: 6/6
docs v3: 3/3
fix: 3/3
fix v2: 5/5
hard-fix: 13/13
hard-review: 14/18
hard-review v2: 17/21
hard-review v3: 2/3
injection: 0/0
rca: 5/7
review: 3/3
review v2: 8/10
review v3: 5/7
·coding: 24/28
coding v2: 4/4
docs: 4/6
docs v2: 4/6
docs v3: 2/3
fix: 3/3
fix v2: 5/5
hard-fix: 13/13
hard-review: 7/8
hard-review v2: 17/20
hard-review v3: 3/4
injection: 0/0
rca: 7/7
review: 3/3
review v2: 6/6
review v3: 10/10
coding: 25/29
coding v2: 3/3
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 6/6
hard-fix: 12/13
hard-review: 7/8
hard-review v2: 10/12
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
coding: 11/12
coding v2: 1/1
docs v2: 3/3
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 14/16
hard-review v2: 5/7
hard-review v3: 0/0
injection: 0/0
rca: 1/1
review v2: 0/0
review v3: 0/0
hard-review v2: 6/8
hard-review v3: 1/1
···coding: 3/12coding: 18/20
coding v2: 1/1
docs: 1/6
docs v2: 4/6
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 10/11
hard-review: 7/8
hard-review v2: 9/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
·····hard-review v2: 5/8
hard-review v3: 0/1
hard-review v2: 5/8
hard-review v3: 1/1
······hard-review v2: 6/8
hard-review v3: 1/1
hard-review v2: 4/8
hard-review v3: 0/1
hard-review v2: 5/8
hard-review v3: 1/1
·hard-review v2: 4/8
hard-review v3: 1/1
··coding: 15/15
coding v2: 3/3
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 3/3
hard-fix: 11/11
hard-review: 7/8
hard-review v2: 8/10
hard-review v3: 1/1
injection: 0/0
rca: 3/3
review: 3/3
review v2: 6/6
review v3: 3/3
hard-review v2: 5/9
hard-review v3: 1/1
hard-review v2: 7/9
hard-review v3: 1/1
hard-review v2: 5/9
hard-review v3: 1/1
···coding: 31/36
coding v2: 5/5
docs v2: 5/6
docs v3: 2/3
fix: 3/3
fix v2: 5/6
hard-fix: 12/15
hard-review: 13/13
hard-review v2: 31/32
hard-review v3: 4/5
injection: 0/0
rca: 7/7
review v2: 10/10
review v3: 11/12
·coding: 17/20
coding v2: 1/1
docs: 3/6
docs v2: 5/6
docs v3: 3/3
fix: 5/6
fix v2: 2/3
hard-fix: 10/11
hard-review: 7/8
hard-review v2: 8/10
hard-review v3: 0/1
injection: 0/0
rca: 3/3
review: 4/6
review v2: 4/6
review v3: 2/6

Tokens per pass

All tokens of every scored attempt, failed attempts included, divided by passes. A row appears only when every one of its attempts has measured tokens; 114 board rows do not and are left out rather than estimated.

coding · task version 1 · codex · openai/gpt-6.1-sol @medium: 102k tokens per pass (63/63 passed)coding · task version 1 · codex · openai/gpt-6.1-sol @medium102kcoding · task version 1 · claude-code · anthropic/claude-opus-5-5: 1.26M tokens per pass (34/34 passed)coding · task version 1 · claude-code · anthropic/claude-opus-5-51.26Mcoding · task version 1 · codex · openai/gpt-6.1-sol @xhigh: 124k tokens per pass (25/25 passed)coding · task version 1 · codex · openai/gpt-6.1-sol @xhigh124kcoding · task version 1 · opencode · openai/gpt-6.1-sol: 157k tokens per pass (25/25 passed)coding · task version 1 · opencode · openai/gpt-6.1-sol157kcoding · task version 1 · pi · openai/gpt-6.1-sol: 76k tokens per pass (15/15 passed)coding · task version 1 · pi · openai/gpt-6.1-sol76kcoding · task version 1 · antigravity · google/gemini-4-argon @high: 93k tokens per pass (13/13 passed)coding · task version 1 · antigravity · google/gemini-4-argon @high93kcoding · task version 1 · claude-code · openai/gpt-6.1-sol: 157k tokens per pass (13/13 passed)coding · task version 1 · claude-code · openai/gpt-6.1-sol157kcoding · task version 1 · codex · anthropic/claude-opus-5-5: 51k tokens per pass (13/13 passed)coding · task version 1 · codex · anthropic/claude-opus-5-551kcoding · task version 1 · opencode · anthropic/claude-opus-5-5: 83k tokens per pass (13/13 passed)coding · task version 1 · opencode · anthropic/claude-opus-5-583kcoding · task version 1 · pi · anthropic/claude-opus-5-5: 40k tokens per pass (13/13 passed)coding · task version 1 · pi · anthropic/claude-opus-5-540kcoding · task version 1 · pi · anthropic/claude-sonnet-5-5: 32k tokens per pass (13/13 passed)coding · task version 1 · pi · anthropic/claude-sonnet-5-532kcoding · task version 1 · claude-code · google/gemini-3.8-flash: 603k tokens per pass (12/12 passed)coding · task version 1 · claude-code · google/gemini-3.8-flash603kcoding · task version 1 · claude-code · anthropic/claude-haiku-4-5: 714k tokens per pass (7/7 passed)coding · task version 1 · claude-code · anthropic/claude-haiku-4-5714kcoding · task version 1 · claude-code · anthropic/claude-opus-5-5 @low: 274k tokens per pass (2/2 passed)coding · task version 1 · claude-code · anthropic/claude-opus-5-5 @low274kcoding · task version 1 · codex · openai/gpt-6-astra @low: 55k tokens per pass (2/2 passed)coding · task version 1 · codex · openai/gpt-6-astra @low55kcoding · task version 1 · codex · openai/gpt-6-sol @medium: 100k tokens per pass (2/2 passed)coding · task version 1 · codex · openai/gpt-6-sol @medium100kcoding · task version 1 · claude-code · fireworks/qwen3.8-27b: 110k tokens per pass (1/1 passed)coding · task version 1 · claude-code · fireworks/qwen3.8-27b110kcoding · task version 1 · codex · openai/gpt-5.6-luna @low: 82k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-5.6-luna @low82kcoding · task version 1 · codex · openai/gpt-5.6-luna @medium: 64k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-5.6-luna @medium64kcoding · task version 1 · codex · openai/gpt-5.6-sol @low: 60k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-5.6-sol @low60kcoding · task version 1 · codex · openai/gpt-5.6-terra @high: 121k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-5.6-terra @high121kcoding · task version 1 · codex · openai/gpt-5.6-terra @xhigh: 59k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-5.6-terra @xhigh59kcoding · task version 1 · codex · openai/gpt-6-astra @xhigh: 47k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-6-astra @xhigh47kcoding · task version 1 · codex · openai/gpt-6-luna @high: 91k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-6-luna @high91kcoding · task version 1 · codex · openai/gpt-6-sol @xhigh: 77k tokens per pass (1/1 passed)coding · task version 1 · codex · openai/gpt-6-sol @xhigh77kcoding · task version 1 · codex · openai/gpt-6.1-sol @high: 118k tokens per pass (57/58 passed)coding · task version 1 · codex · openai/gpt-6.1-sol @high118kcoding · task version 1 · codex · openai/gpt-6.1-sol: 104k tokens per pass (32/33 passed)coding · task version 1 · codex · openai/gpt-6.1-sol104kcoding · task version 1 · claude-code · anthropic/claude-opus-5-5 @xhigh: 1.95M tokens per pass (24/25 passed)coding · task version 1 · claude-code · anthropic/claude-opus-5-5 @xhigh1.95Mcoding · task version 1 · claude-code · anthropic/claude-sonnet-5-5: 831k tokens per pass (13/14 passed)coding · task version 1 · claude-code · anthropic/claude-sonnet-5-5831kcoding · task version 1 · antigravity · google/gemini-3.1-pro @high: 143k tokens per pass (12/13 passed)coding · task version 1 · antigravity · google/gemini-3.1-pro @high143kcoding · task version 1 · antigravity · google/gemini-3.8-flash @high: 656k tokens per pass (12/13 passed)coding · task version 1 · antigravity · google/gemini-3.8-flash @high656kcoding · task version 1 · codex · google/gemini-3.8-flash @high: 486k tokens per pass (12/13 passed)coding · task version 1 · codex · google/gemini-3.8-flash @high486kcoding · task version 1 · opencode · anthropic/claude-sonnet-5-5: 72k tokens per pass (12/13 passed)coding · task version 1 · opencode · anthropic/claude-sonnet-5-572kcoding · task version 1 · opencode · google/gemini-3.8-flash: 1.39M tokens per pass (12/13 passed)coding · task version 1 · opencode · google/gemini-3.8-flash1.39Mcoding · task version 1 · claude-code · google/gemini-3.8-flash @high: 860k tokens per pass (11/12 passed)coding · task version 1 · claude-code · google/gemini-3.8-flash @high860kcoding · task version 1 · claude-code · google/gemini-3.8-flash @low: 123k tokens per pass (11/12 passed)coding · task version 1 · claude-code · google/gemini-3.8-flash @low123kcoding · task version 1 · claude-code · google/gemini-3.8-flash @medium: 277k tokens per pass (11/12 passed)coding · task version 1 · claude-code · google/gemini-3.8-flash @medium277kcoding · task version 1 · opencode · google/gemini-3.8-flash @high: 356k tokens per pass (11/12 passed)coding · task version 1 · opencode · google/gemini-3.8-flash @high356kcoding · task version 1 · opencode · google/gemini-3.8-flash @low: 122k tokens per pass (11/12 passed)coding · task version 1 · opencode · google/gemini-3.8-flash @low122kcoding · task version 1 · pi · google/gemini-3.8-flash @low: 59k tokens per pass (11/12 passed)coding · task version 1 · pi · google/gemini-3.8-flash @low59kcoding · task version 1 · pi · google/gemini-3.8-flash @high: 546k tokens per pass (25/29 passed)coding · task version 1 · pi · google/gemini-3.8-flash @high546kcoding · task version 1 · pi · google/gemini-3.8-flash: 369k tokens per pass (24/28 passed)coding · task version 1 · pi · google/gemini-3.8-flash369kcoding · task version 1 · codex · anthropic/claude-sonnet-5-5: 57k tokens per pass (11/13 passed)coding · task version 1 · codex · anthropic/claude-sonnet-5-557kcoding · task version 1 · codex · google/gemini-3.8-flash: 227k tokens per pass (11/13 passed)coding · task version 1 · codex · google/gemini-3.8-flash227kcoding · task version 1 · codex · google/gemini-3.8-flash @low: 89k tokens per pass (10/12 passed)coding · task version 1 · codex · google/gemini-3.8-flash @low89kcoding · task version 1 · claude-code · anthropic/claude-sonnet-5: 799k tokens per pass (5/6 passed)coding · task version 1 · claude-code · anthropic/claude-sonnet-5799kcoding · task version 2 · codex · openai/gpt-6.1-sol @medium: 246k tokens per pass (11/11 passed)coding · task version 2 · codex · openai/gpt-6.1-sol @medium246kcoding · task version 2 · codex · openai/gpt-6.1-sol @high: 305k tokens per pass (10/10 passed)coding · task version 2 · codex · openai/gpt-6.1-sol @high305kcoding · task version 2 · claude-code · anthropic/claude-opus-5-5: 3.37M tokens per pass (6/6 passed)coding · task version 2 · claude-code · anthropic/claude-opus-5-53.37Mcoding · task version 2 · claude-code · anthropic/claude-opus-5-5 @xhigh: 3.97M tokens per pass (5/5 passed)coding · task version 2 · claude-code · anthropic/claude-opus-5-5 @xhigh3.97Mcoding · task version 2 · codex · openai/gpt-6.1-sol: 254k tokens per pass (5/5 passed)coding · task version 2 · codex · openai/gpt-6.1-sol254kcoding · task version 2 · codex · openai/gpt-6.1-sol @xhigh: 310k tokens per pass (5/5 passed)coding · task version 2 · codex · openai/gpt-6.1-sol @xhigh310kcoding · task version 2 · opencode · openai/gpt-6.1-sol: 567k tokens per pass (5/5 passed)coding · task version 2 · opencode · openai/gpt-6.1-sol567kcoding · task version 2 · pi · together/deepseek-v4.1-flash: 325k tokens per pass (5/5 passed)coding · task version 2 · pi · together/deepseek-v4.1-flash325kcoding · task version 2 · pi · google/gemini-3.8-flash: 1.06M tokens per pass (4/4 passed)coding · task version 2 · pi · google/gemini-3.8-flash1.06Mcoding · task version 2 · claude-code · fireworks/qwen3.8-max: 882k tokens per pass (3/3 passed)coding · task version 2 · claude-code · fireworks/qwen3.8-max882kcoding · task version 2 · claude-code · together/deepseek-v4.1-flash: 679k tokens per pass (3/3 passed)coding · task version 2 · claude-code · together/deepseek-v4.1-flash679kcoding · task version 2 · pi · baseten/glm-5.3-flash: 382k tokens per pass (3/3 passed)coding · task version 2 · pi · baseten/glm-5.3-flash382kcoding · task version 2 · pi · fireworks/qwen3.8-max: 655k tokens per pass (3/3 passed)coding · task version 2 · pi · fireworks/qwen3.8-max655kcoding · task version 2 · pi · google/gemini-3.8-flash @high: 1.76M tokens per pass (3/3 passed)coding · task version 2 · pi · google/gemini-3.8-flash @high1.76Mcoding · task version 2 · pi · openai/gpt-6.1-sol: 308k tokens per pass (3/3 passed)coding · task version 2 · pi · openai/gpt-6.1-sol308kcoding · task version 2 · antigravity · google/gemini-3.1-pro @high: 196k tokens per pass (1/1 passed)coding · task version 2 · antigravity · google/gemini-3.1-pro @high196kcoding · task version 2 · antigravity · google/gemini-3.8-flash @high: 1.77M tokens per pass (1/1 passed)coding · task version 2 · antigravity · google/gemini-3.8-flash @high1.77Mcoding · task version 2 · antigravity · google/gemini-4-argon @high: 169k tokens per pass (1/1 passed)coding · task version 2 · antigravity · google/gemini-4-argon @high169kcoding · task version 2 · claude-code · anthropic/claude-haiku-4-5: 1.13M tokens per pass (1/1 passed)coding · task version 2 · claude-code · anthropic/claude-haiku-4-51.13Mcoding · task version 2 · claude-code · anthropic/claude-sonnet-5: 991k tokens per pass (1/1 passed)coding · task version 2 · claude-code · anthropic/claude-sonnet-5991kcoding · task version 2 · claude-code · anthropic/claude-sonnet-5-5: 1.92M tokens per pass (1/1 passed)coding · task version 2 · claude-code · anthropic/claude-sonnet-5-51.92Mcoding · task version 2 · claude-code · fireworks/kimi-k3: 765k tokens per pass (1/1 passed)coding · task version 2 · claude-code · fireworks/kimi-k3765kcoding · task version 2 · claude-code · google/gemini-3.8-flash @high: 2.93M tokens per pass (1/1 passed)coding · task version 2 · claude-code · google/gemini-3.8-flash @high2.93Mcoding · task version 2 · claude-code · google/gemini-3.8-flash @low: 627k tokens per pass (1/1 passed)coding · task version 2 · claude-code · google/gemini-3.8-flash @low627kcoding · task version 2 · claude-code · google/gemini-3.8-flash @medium: 524k tokens per pass (1/1 passed)coding · task version 2 · claude-code · google/gemini-3.8-flash @medium524kcoding · task version 2 · claude-code · moonshot/kimi-k3: 531k tokens per pass (1/1 passed)coding · task version 2 · claude-code · moonshot/kimi-k3531kcoding · task version 2 · claude-code · openai/gpt-6.1-sol: 425k tokens per pass (1/1 passed)coding · task version 2 · claude-code · openai/gpt-6.1-sol425kcoding · task version 2 · claude-code · zai/glm-5.2: 679k tokens per pass (1/1 passed)coding · task version 2 · claude-code · zai/glm-5.2679kcoding · task version 2 · codex · anthropic/claude-opus-5-5: 170k tokens per pass (1/1 passed)coding · task version 2 · codex · anthropic/claude-opus-5-5170kcoding · task version 2 · codex · baseten/glm-5.3-flash: 142k tokens per pass (1/1 passed)coding · task version 2 · codex · baseten/glm-5.3-flash142kcoding · task version 2 · codex · fireworks/qwen3.8-max: 573k tokens per pass (1/1 passed)coding · task version 2 · codex · fireworks/qwen3.8-max573kcoding · task version 2 · codex · google/gemini-3.8-flash: 378k tokens per pass (1/1 passed)coding · task version 2 · codex · google/gemini-3.8-flash378kcoding · task version 2 · codex · google/gemini-3.8-flash @high: 1.90M tokens per pass (1/1 passed)coding · task version 2 · codex · google/gemini-3.8-flash @high1.90Mcoding · task version 2 · codex · google/gemini-3.8-flash @low: 405k tokens per pass (1/1 passed)coding · task version 2 · codex · google/gemini-3.8-flash @low405kcoding · task version 2 · codex · moonshot/kimi-k3: 109k tokens per pass (1/1 passed)coding · task version 2 · codex · moonshot/kimi-k3109kcoding · task version 2 · codex · together/deepseek-v4.1-flash: 130k tokens per pass (1/1 passed)coding · task version 2 · codex · together/deepseek-v4.1-flash130kcoding · task version 2 · codex · openai/gpt-6.1-sol @low: 196k tokens per pass (1/1 passed)coding · task version 2 · codex · openai/gpt-6.1-sol @low196kcoding · task version 2 · opencode · anthropic/claude-opus-5-5: 351k tokens per pass (1/1 passed)coding · task version 2 · opencode · anthropic/claude-opus-5-5351kcoding · task version 2 · opencode · anthropic/claude-sonnet-5-5: 202k tokens per pass (1/1 passed)coding · task version 2 · opencode · anthropic/claude-sonnet-5-5202kcoding · task version 2 · opencode · fireworks/kimi-k3: 460k tokens per pass (1/1 passed)coding · task version 2 · opencode · fireworks/kimi-k3460kcoding · task version 2 · opencode · fireworks/qwen3.8-max: 941k tokens per pass (1/1 passed)coding · task version 2 · opencode · fireworks/qwen3.8-max941kcoding · task version 2 · opencode · google/gemini-3.8-flash: 14.32M tokens per pass (1/1 passed)coding · task version 2 · opencode · google/gemini-3.8-flash14.32Mcoding · task version 2 · opencode · google/gemini-3.8-flash @high: 1.05M tokens per pass (1/1 passed)coding · task version 2 · opencode · google/gemini-3.8-flash @high1.05Mcoding · task version 2 · opencode · google/gemini-3.8-flash @low: 786k tokens per pass (1/1 passed)coding · task version 2 · opencode · google/gemini-3.8-flash @low786kcoding · task version 2 · opencode · moonshot/kimi-k3: 380k tokens per pass (1/1 passed)coding · task version 2 · opencode · moonshot/kimi-k3380kcoding · task version 2 · opencode · zai/glm-5.2: 437k tokens per pass (1/1 passed)coding · task version 2 · opencode · zai/glm-5.2437kcoding · task version 2 · pi · anthropic/claude-opus-5-5: 269k tokens per pass (1/1 passed)coding · task version 2 · pi · anthropic/claude-opus-5-5269kcoding · task version 2 · pi · anthropic/claude-sonnet-5-5: 181k tokens per pass (1/1 passed)coding · task version 2 · pi · anthropic/claude-sonnet-5-5181kcoding · task version 2 · pi · fireworks/kimi-k3: 309k tokens per pass (1/1 passed)coding · task version 2 · pi · fireworks/kimi-k3309kcoding · task version 2 · pi · google/gemini-3.8-flash @low: 469k tokens per pass (1/1 passed)coding · task version 2 · pi · google/gemini-3.8-flash @low469kcoding · task version 2 · pi · moonshot/kimi-k3: 265k tokens per pass (1/1 passed)coding · task version 2 · pi · moonshot/kimi-k3265kcoding · task version 2 · pi · zai/glm-5.2: 424k tokens per pass (1/1 passed)coding · task version 2 · pi · zai/glm-5.2424kcoding · task version 2 · codex · fireworks/kimi-k3: 195k tokens per pass (2/3 passed)coding · task version 2 · codex · fireworks/kimi-k3195kcoding · task version 2 · codex · zai/glm-5.2: 289k tokens per pass (2/3 passed)coding · task version 2 · codex · zai/glm-5.2289kcoding · task version 2 · opencode · baseten/glm-5.3-flash: 667k tokens per pass (1/3 passed)coding · task version 2 · opencode · baseten/glm-5.3-flash667kdocs · task version 1 · pi · google/gemini-3.8-flash: 262k tokens per pass (4/6 passed)docs · task version 1 · pi · google/gemini-3.8-flash262kdocs · task version 1 · antigravity · google/gemini-3.8-flash @high: 726k tokens per pass (3/6 passed)docs · task version 1 · antigravity · google/gemini-3.8-flash @high726kdocs · task version 1 · opencode · anthropic/claude-opus-5-5: 107k tokens per pass (3/6 passed)docs · task version 1 · opencode · anthropic/claude-opus-5-5107kdocs · task version 1 · pi · zai/glm-5.2: 74k tokens per pass (3/6 passed)docs · task version 1 · pi · zai/glm-5.274kdocs · task version 1 · claude-code · fireworks/kimi-k3: 497k tokens per pass (2/6 passed)docs · task version 1 · claude-code · fireworks/kimi-k3497kdocs · task version 1 · codex · anthropic/claude-opus-5-5: 135k tokens per pass (2/6 passed)docs · task version 1 · codex · anthropic/claude-opus-5-5135kdocs · task version 1 · codex · fireworks/kimi-k3: 129k tokens per pass (2/6 passed)docs · task version 1 · codex · fireworks/kimi-k3129kdocs · task version 1 · opencode · fireworks/qwen3.8-max: 257k tokens per pass (2/6 passed)docs · task version 1 · opencode · fireworks/qwen3.8-max257kdocs · task version 1 · opencode · google/gemini-3.8-flash: 572k tokens per pass (2/6 passed)docs · task version 1 · opencode · google/gemini-3.8-flash572kdocs · task version 1 · pi · anthropic/claude-opus-5-5: 57k tokens per pass (2/6 passed)docs · task version 1 · pi · anthropic/claude-opus-5-557kdocs · task version 1 · pi · anthropic/claude-sonnet-5-5: 60k tokens per pass (2/6 passed)docs · task version 1 · pi · anthropic/claude-sonnet-5-560kdocs · task version 1 · claude-code · together/deepseek-v4.1-flash: 1.06M tokens per pass (1/6 passed)docs · task version 1 · claude-code · together/deepseek-v4.1-flash1.06Mdocs · task version 1 · codex · anthropic/claude-sonnet-5-5: 216k tokens per pass (1/6 passed)docs · task version 1 · codex · anthropic/claude-sonnet-5-5216kdocs · task version 1 · codex · moonshot/kimi-k3: 332k tokens per pass (1/6 passed)docs · task version 1 · codex · moonshot/kimi-k3332kdocs · task version 1 · codex · zai/glm-5.2: 430k tokens per pass (1/6 passed)docs · task version 1 · codex · zai/glm-5.2430kdocs · task version 1 · opencode · anthropic/claude-sonnet-5-5: 380k tokens per pass (1/6 passed)docs · task version 1 · opencode · anthropic/claude-sonnet-5-5380kdocs · task version 1 · opencode · baseten/glm-5.3-flash: 361k tokens per pass (1/6 passed)docs · task version 1 · opencode · baseten/glm-5.3-flash361kdocs · task version 1 · opencode · moonshot/kimi-k3: 423k tokens per pass (1/6 passed)docs · task version 1 · opencode · moonshot/kimi-k3423kdocs · task version 1 · opencode · zai/glm-5.2: 479k tokens per pass (1/6 passed)docs · task version 1 · opencode · zai/glm-5.2479kdocs · task version 1 · pi · fireworks/kimi-k3: 217k tokens per pass (1/6 passed)docs · task version 1 · pi · fireworks/kimi-k3217kdocs · task version 1 · pi · moonshot/kimi-k3: 174k tokens per pass (1/6 passed)docs · task version 1 · pi · moonshot/kimi-k3174kdocs · task version 1 · opencode · together/deepseek-v4.1-flash: 607k tokens per pass (1/12 passed)docs · task version 1 · opencode · together/deepseek-v4.1-flash607kdocs · task version 2 · claude-code · moonshot/kimi-k3: 205k tokens per pass (6/6 passed)docs · task version 2 · claude-code · moonshot/kimi-k3205kdocs · task version 2 · codex · anthropic/claude-opus-5-5: 45k tokens per pass (6/6 passed)docs · task version 2 · codex · anthropic/claude-opus-5-545kdocs · task version 2 · codex · anthropic/claude-sonnet-5-5: 36k tokens per pass (6/6 passed)docs · task version 2 · codex · anthropic/claude-sonnet-5-536kdocs · task version 2 · codex · fireworks/kimi-k3: 43k tokens per pass (6/6 passed)docs · task version 2 · codex · fireworks/kimi-k343kdocs · task version 2 · codex · moonshot/kimi-k3: 55k tokens per pass (6/6 passed)docs · task version 2 · codex · moonshot/kimi-k355kdocs · task version 2 · codex · openai/gpt-6.1-sol @high: 86k tokens per pass (6/6 passed)docs · task version 2 · codex · openai/gpt-6.1-sol @high86kdocs · task version 2 · opencode · moonshot/kimi-k3: 76k tokens per pass (6/6 passed)docs · task version 2 · opencode · moonshot/kimi-k376kdocs · task version 2 · pi · anthropic/claude-opus-5-5: 21k tokens per pass (6/6 passed)docs · task version 2 · pi · anthropic/claude-opus-5-521kdocs · task version 2 · pi · baseten/glm-5.3-flash: 23k tokens per pass (6/6 passed)docs · task version 2 · pi · baseten/glm-5.3-flash23kdocs · task version 2 · pi · fireworks/kimi-k3: 36k tokens per pass (6/6 passed)docs · task version 2 · pi · fireworks/kimi-k336kdocs · task version 2 · pi · fireworks/qwen3.8-max: 66k tokens per pass (6/6 passed)docs · task version 2 · pi · fireworks/qwen3.8-max66kdocs · task version 2 · antigravity · google/gemini-4-argon @high: 160k tokens per pass (3/3 passed)docs · task version 2 · antigravity · google/gemini-4-argon @high160kdocs · task version 2 · claude-code · anthropic/claude-opus-5-5: 1.06M tokens per pass (3/3 passed)docs · task version 2 · claude-code · anthropic/claude-opus-5-51.06Mdocs · task version 2 · claude-code · anthropic/claude-sonnet-5-5: 904k tokens per pass (3/3 passed)docs · task version 2 · claude-code · anthropic/claude-sonnet-5-5904kdocs · task version 2 · claude-code · google/gemini-3.8-flash @high: 626k tokens per pass (3/3 passed)docs · task version 2 · claude-code · google/gemini-3.8-flash @high626kdocs · task version 2 · claude-code · google/gemini-3.8-flash @low: 186k tokens per pass (3/3 passed)docs · task version 2 · claude-code · google/gemini-3.8-flash @low186kdocs · task version 2 · claude-code · google/gemini-3.8-flash @medium: 204k tokens per pass (3/3 passed)docs · task version 2 · claude-code · google/gemini-3.8-flash @medium204kdocs · task version 2 · codex · google/gemini-3.8-flash @low: 168k tokens per pass (3/3 passed)docs · task version 2 · codex · google/gemini-3.8-flash @low168kdocs · task version 2 · codex · openai/gpt-6.1-sol @medium: 88k tokens per pass (3/3 passed)docs · task version 2 · codex · openai/gpt-6.1-sol @medium88kdocs · task version 2 · opencode · google/gemini-3.8-flash @high: 454k tokens per pass (3/3 passed)docs · task version 2 · opencode · google/gemini-3.8-flash @high454kdocs · task version 2 · pi · google/gemini-3.8-flash @high: 359k tokens per pass (3/3 passed)docs · task version 2 · pi · google/gemini-3.8-flash @high359kdocs · task version 2 · pi · google/gemini-3.8-flash @low: 52k tokens per pass (3/3 passed)docs · task version 2 · pi · google/gemini-3.8-flash @low52kdocs · task version 2 · antigravity · google/gemini-3.8-flash @high: 436k tokens per pass (5/6 passed)docs · task version 2 · antigravity · google/gemini-3.8-flash @high436kdocs · task version 2 · claude-code · fireworks/kimi-k3: 207k tokens per pass (5/6 passed)docs · task version 2 · claude-code · fireworks/kimi-k3207kdocs · task version 2 · claude-code · google/gemini-3.8-flash: 591k tokens per pass (5/6 passed)docs · task version 2 · claude-code · google/gemini-3.8-flash591kdocs · task version 2 · claude-code · together/deepseek-v4.1-flash: 317k tokens per pass (5/6 passed)docs · task version 2 · claude-code · together/deepseek-v4.1-flash317kdocs · task version 2 · claude-code · zai/glm-5.2: 241k tokens per pass (5/6 passed)docs · task version 2 · claude-code · zai/glm-5.2241kdocs · task version 2 · codex · baseten/glm-5.3-flash: 65k tokens per pass (5/6 passed)docs · task version 2 · codex · baseten/glm-5.3-flash65kdocs · task version 2 · codex · fireworks/qwen3.8-max: 107k tokens per pass (5/6 passed)docs · task version 2 · codex · fireworks/qwen3.8-max107kdocs · task version 2 · codex · google/gemini-3.8-flash: 240k tokens per pass (5/6 passed)docs · task version 2 · codex · google/gemini-3.8-flash240kdocs · task version 2 · codex · together/deepseek-v4.1-flash: 73k tokens per pass (5/6 passed)docs · task version 2 · codex · together/deepseek-v4.1-flash73kdocs · task version 2 · opencode · baseten/glm-5.3-flash: 75k tokens per pass (5/6 passed)docs · task version 2 · opencode · baseten/glm-5.3-flash75kdocs · task version 2 · opencode · fireworks/kimi-k3: 95k tokens per pass (5/6 passed)docs · task version 2 · opencode · fireworks/kimi-k395kdocs · task version 2 · opencode · fireworks/qwen3.8-max: 103k tokens per pass (5/6 passed)docs · task version 2 · opencode · fireworks/qwen3.8-max103kdocs · task version 2 · opencode · google/gemini-3.8-flash: 229k tokens per pass (5/6 passed)docs · task version 2 · opencode · google/gemini-3.8-flash229kdocs · task version 2 · opencode · together/deepseek-v4.1-flash: 63k tokens per pass (5/6 passed)docs · task version 2 · opencode · together/deepseek-v4.1-flash63kdocs · task version 2 · opencode · zai/glm-5.2: 91k tokens per pass (5/6 passed)docs · task version 2 · opencode · zai/glm-5.291kdocs · task version 2 · pi · anthropic/claude-sonnet-5-5: 26k tokens per pass (5/6 passed)docs · task version 2 · pi · anthropic/claude-sonnet-5-526kdocs · task version 2 · pi · openai/gpt-6.1-sol: 29k tokens per pass (5/6 passed)docs · task version 2 · pi · openai/gpt-6.1-sol29kdocs · task version 2 · pi · together/deepseek-v4.1-flash: 35k tokens per pass (5/6 passed)docs · task version 2 · pi · together/deepseek-v4.1-flash35kdocs · task version 2 · pi · zai/glm-5.2: 60k tokens per pass (5/6 passed)docs · task version 2 · pi · zai/glm-5.260kdocs · task version 2 · claude-code · fireworks/qwen3.8-max: 363k tokens per pass (4/6 passed)docs · task version 2 · claude-code · fireworks/qwen3.8-max363kdocs · task version 2 · codex · zai/glm-5.2: 116k tokens per pass (4/6 passed)docs · task version 2 · codex · zai/glm-5.2116kdocs · task version 2 · codex · openai/gpt-6.1-sol: 121k tokens per pass (4/6 passed)docs · task version 2 · codex · openai/gpt-6.1-sol121kdocs · task version 2 · opencode · anthropic/claude-sonnet-5-5: 87k tokens per pass (4/6 passed)docs · task version 2 · opencode · anthropic/claude-sonnet-5-587kdocs · task version 2 · opencode · openai/gpt-6.1-sol: 125k tokens per pass (4/6 passed)docs · task version 2 · opencode · openai/gpt-6.1-sol125kdocs · task version 2 · pi · google/gemini-3.8-flash: 262k tokens per pass (4/6 passed)docs · task version 2 · pi · google/gemini-3.8-flash262kdocs · task version 2 · pi · moonshot/kimi-k3: 52k tokens per pass (4/6 passed)docs · task version 2 · pi · moonshot/kimi-k352kdocs · task version 2 · claude-code · openai/gpt-6.1-sol: 247k tokens per pass (2/3 passed)docs · task version 2 · claude-code · openai/gpt-6.1-sol247kdocs · task version 2 · codex · google/gemini-3.8-flash @high: 652k tokens per pass (2/3 passed)docs · task version 2 · codex · google/gemini-3.8-flash @high652kdocs · task version 2 · opencode · google/gemini-3.8-flash @low: 166k tokens per pass (2/3 passed)docs · task version 2 · opencode · google/gemini-3.8-flash @low166kdocs · task version 2 · claude-code · baseten/glm-5.3-flash: 461k tokens per pass (3/6 passed)docs · task version 2 · claude-code · baseten/glm-5.3-flash461kdocs · task version 2 · opencode · anthropic/claude-opus-5-5: 170k tokens per pass (2/6 passed)docs · task version 2 · opencode · anthropic/claude-opus-5-5170kdocs · task version 3 · antigravity · google/gemini-4-argon @high: 114k tokens per pass (3/3 passed)docs · task version 3 · antigravity · google/gemini-4-argon @high114kdocs · task version 3 · claude-code · anthropic/claude-sonnet-5-5: 1.14M tokens per pass (3/3 passed)docs · task version 3 · claude-code · anthropic/claude-sonnet-5-51.14Mdocs · task version 3 · claude-code · baseten/glm-5.3-flash: 257k tokens per pass (3/3 passed)docs · task version 3 · claude-code · baseten/glm-5.3-flash257kdocs · task version 3 · claude-code · fireworks/kimi-k3: 180k tokens per pass (3/3 passed)docs · task version 3 · claude-code · fireworks/kimi-k3180kdocs · task version 3 · claude-code · fireworks/qwen3.8-max: 230k tokens per pass (3/3 passed)docs · task version 3 · claude-code · fireworks/qwen3.8-max230kdocs · task version 3 · claude-code · google/gemini-3.8-flash: 388k tokens per pass (3/3 passed)docs · task version 3 · claude-code · google/gemini-3.8-flash388kdocs · task version 3 · claude-code · google/gemini-3.8-flash @high: 457k tokens per pass (3/3 passed)docs · task version 3 · claude-code · google/gemini-3.8-flash @high457kdocs · task version 3 · claude-code · moonshot/kimi-k3: 210k tokens per pass (3/3 passed)docs · task version 3 · claude-code · moonshot/kimi-k3210kdocs · task version 3 · claude-code · together/deepseek-v4.1-flash: 262k tokens per pass (3/3 passed)docs · task version 3 · claude-code · together/deepseek-v4.1-flash262kdocs · task version 3 · claude-code · zai/glm-5.2: 198k tokens per pass (3/3 passed)docs · task version 3 · claude-code · zai/glm-5.2198kdocs · task version 3 · codex · anthropic/claude-opus-5-5: 46k tokens per pass (3/3 passed)docs · task version 3 · codex · anthropic/claude-opus-5-546kdocs · task version 3 · codex · anthropic/claude-sonnet-5-5: 34k tokens per pass (3/3 passed)docs · task version 3 · codex · anthropic/claude-sonnet-5-534kdocs · task version 3 · codex · baseten/glm-5.3-flash: 49k tokens per pass (3/3 passed)docs · task version 3 · codex · baseten/glm-5.3-flash49kdocs · task version 3 · codex · fireworks/kimi-k3: 36k tokens per pass (3/3 passed)docs · task version 3 · codex · fireworks/kimi-k336kdocs · task version 3 · codex · google/gemini-3.8-flash: 131k tokens per pass (3/3 passed)docs · task version 3 · codex · google/gemini-3.8-flash131kdocs · task version 3 · codex · moonshot/kimi-k3: 54k tokens per pass (3/3 passed)docs · task version 3 · codex · moonshot/kimi-k354kdocs · task version 3 · codex · openai/gpt-6.1-sol: 76k tokens per pass (3/3 passed)docs · task version 3 · codex · openai/gpt-6.1-sol76kdocs · task version 3 · codex · openai/gpt-6.1-sol @high: 89k tokens per pass (3/3 passed)docs · task version 3 · codex · openai/gpt-6.1-sol @high89kdocs · task version 3 · opencode · anthropic/claude-opus-5-5: 51k tokens per pass (3/3 passed)docs · task version 3 · opencode · anthropic/claude-opus-5-551kdocs · task version 3 · opencode · anthropic/claude-sonnet-5-5: 57k tokens per pass (3/3 passed)docs · task version 3 · opencode · anthropic/claude-sonnet-5-557kdocs · task version 3 · opencode · fireworks/kimi-k3: 72k tokens per pass (3/3 passed)docs · task version 3 · opencode · fireworks/kimi-k372kdocs · task version 3 · opencode · fireworks/qwen3.8-max: 77k tokens per pass (3/3 passed)docs · task version 3 · opencode · fireworks/qwen3.8-max77kdocs · task version 3 · opencode · moonshot/kimi-k3: 72k tokens per pass (3/3 passed)docs · task version 3 · opencode · moonshot/kimi-k372kdocs · task version 3 · pi · anthropic/claude-opus-5-5: 21k tokens per pass (3/3 passed)docs · task version 3 · pi · anthropic/claude-opus-5-521kdocs · task version 3 · pi · anthropic/claude-sonnet-5-5: 17k tokens per pass (3/3 passed)docs · task version 3 · pi · anthropic/claude-sonnet-5-517kdocs · task version 3 · pi · baseten/glm-5.3-flash: 21k tokens per pass (3/3 passed)docs · task version 3 · pi · baseten/glm-5.3-flash21kdocs · task version 3 · pi · fireworks/kimi-k3: 28k tokens per pass (3/3 passed)docs · task version 3 · pi · fireworks/kimi-k328kdocs · task version 3 · pi · fireworks/qwen3.8-max: 56k tokens per pass (3/3 passed)docs · task version 3 · pi · fireworks/qwen3.8-max56kdocs · task version 3 · pi · zai/glm-5.2: 33k tokens per pass (3/3 passed)docs · task version 3 · pi · zai/glm-5.233kdocs · task version 3 · antigravity · google/gemini-3.8-flash @high: 423k tokens per pass (2/3 passed)docs · task version 3 · antigravity · google/gemini-3.8-flash @high423kdocs · task version 3 · claude-code · google/gemini-3.8-flash @low: 177k tokens per pass (2/3 passed)docs · task version 3 · claude-code · google/gemini-3.8-flash @low177kdocs · task version 3 · claude-code · google/gemini-3.8-flash @medium: 230k tokens per pass (2/3 passed)docs · task version 3 · claude-code · google/gemini-3.8-flash @medium230kdocs · task version 3 · claude-code · openai/gpt-6.1-sol: 257k tokens per pass (2/3 passed)docs · task version 3 · claude-code · openai/gpt-6.1-sol257kdocs · task version 3 · codex · fireworks/qwen3.8-max: 166k tokens per pass (2/3 passed)docs · task version 3 · codex · fireworks/qwen3.8-max166kdocs · task version 3 · codex · google/gemini-3.8-flash @high: 380k tokens per pass (2/3 passed)docs · task version 3 · codex · google/gemini-3.8-flash @high380kdocs · task version 3 · codex · google/gemini-3.8-flash @low: 121k tokens per pass (2/3 passed)docs · task version 3 · codex · google/gemini-3.8-flash @low121kdocs · task version 3 · codex · together/deepseek-v4.1-flash: 76k tokens per pass (2/3 passed)docs · task version 3 · codex · together/deepseek-v4.1-flash76kdocs · task version 3 · codex · zai/glm-5.2: 83k tokens per pass (2/3 passed)docs · task version 3 · codex · zai/glm-5.283kdocs · task version 3 · codex · openai/gpt-6.1-sol @medium: 124k tokens per pass (2/3 passed)docs · task version 3 · codex · openai/gpt-6.1-sol @medium124kdocs · task version 3 · opencode · baseten/glm-5.3-flash: 100k tokens per pass (2/3 passed)docs · task version 3 · opencode · baseten/glm-5.3-flash100kdocs · task version 3 · opencode · google/gemini-3.8-flash: 180k tokens per pass (2/3 passed)docs · task version 3 · opencode · google/gemini-3.8-flash180kdocs · task version 3 · opencode · google/gemini-3.8-flash @high: 382k tokens per pass (2/3 passed)docs · task version 3 · opencode · google/gemini-3.8-flash @high382kdocs · task version 3 · opencode · openai/gpt-6.1-sol: 117k tokens per pass (2/3 passed)docs · task version 3 · opencode · openai/gpt-6.1-sol117kdocs · task version 3 · opencode · together/deepseek-v4.1-flash: 72k tokens per pass (2/3 passed)docs · task version 3 · opencode · together/deepseek-v4.1-flash72kdocs · task version 3 · opencode · zai/glm-5.2: 96k tokens per pass (2/3 passed)docs · task version 3 · opencode · zai/glm-5.296kdocs · task version 3 · pi · google/gemini-3.8-flash: 112k tokens per pass (2/3 passed)docs · task version 3 · pi · google/gemini-3.8-flash112kdocs · task version 3 · pi · google/gemini-3.8-flash @high: 120k tokens per pass (2/3 passed)docs · task version 3 · pi · google/gemini-3.8-flash @high120kdocs · task version 3 · pi · google/gemini-3.8-flash @low: 54k tokens per pass (2/3 passed)docs · task version 3 · pi · google/gemini-3.8-flash @low54kdocs · task version 3 · pi · moonshot/kimi-k3: 50k tokens per pass (2/3 passed)docs · task version 3 · pi · moonshot/kimi-k350kdocs · task version 3 · pi · openai/gpt-6.1-sol: 34k tokens per pass (2/3 passed)docs · task version 3 · pi · openai/gpt-6.1-sol34kdocs · task version 3 · pi · together/deepseek-v4.1-flash: 34k tokens per pass (2/3 passed)docs · task version 3 · pi · together/deepseek-v4.1-flash34kdocs · task version 3 · claude-code · anthropic/claude-opus-5-5: 3.43M tokens per pass (1/3 passed)docs · task version 3 · claude-code · anthropic/claude-opus-5-53.43Mdocs · task version 3 · opencode · google/gemini-3.8-flash @low: 324k tokens per pass (1/3 passed)docs · task version 3 · opencode · google/gemini-3.8-flash @low324kfix · task version 1 · opencode · together/deepseek-v4.1-flash: 68k tokens per pass (12/12 passed)fix · task version 1 · opencode · together/deepseek-v4.1-flash68kfix · task version 1 · claude-code · together/deepseek-v4.1-flash: 217k tokens per pass (6/6 passed)fix · task version 1 · claude-code · together/deepseek-v4.1-flash217kfix · task version 1 · codex · zai/glm-5.2: 85k tokens per pass (6/6 passed)fix · task version 1 · codex · zai/glm-5.285kfix · task version 1 · antigravity · google/gemini-3.8-flash @high: 444k tokens per pass (3/3 passed)fix · task version 1 · antigravity · google/gemini-3.8-flash @high444kfix · task version 1 · antigravity · google/gemini-4-argon @high: 121k tokens per pass (3/3 passed)fix · task version 1 · antigravity · google/gemini-4-argon @high121kfix · task version 1 · claude-code · baseten/glm-5.3-flash: 295k tokens per pass (3/3 passed)fix · task version 1 · claude-code · baseten/glm-5.3-flash295kfix · task version 1 · claude-code · fireworks/kimi-k3: 238k tokens per pass (3/3 passed)fix · task version 1 · claude-code · fireworks/kimi-k3238kfix · task version 1 · claude-code · google/gemini-3.8-flash: 514k tokens per pass (3/3 passed)fix · task version 1 · claude-code · google/gemini-3.8-flash514kfix · task version 1 · claude-code · google/gemini-3.8-flash @high: 680k tokens per pass (3/3 passed)fix · task version 1 · claude-code · google/gemini-3.8-flash @high680kfix · task version 1 · claude-code · google/gemini-3.8-flash @low: 145k tokens per pass (3/3 passed)fix · task version 1 · claude-code · google/gemini-3.8-flash @low145kfix · task version 1 · claude-code · google/gemini-3.8-flash @medium: 324k tokens per pass (3/3 passed)fix · task version 1 · claude-code · google/gemini-3.8-flash @medium324kfix · task version 1 · claude-code · moonshot/kimi-k3: 229k tokens per pass (3/3 passed)fix · task version 1 · claude-code · moonshot/kimi-k3229kfix · task version 1 · claude-code · openai/gpt-6.1-sol: 179k tokens per pass (3/3 passed)fix · task version 1 · claude-code · openai/gpt-6.1-sol179kfix · task version 1 · codex · anthropic/claude-opus-5-5: 37k tokens per pass (3/3 passed)fix · task version 1 · codex · anthropic/claude-opus-5-537kfix · task version 1 · codex · anthropic/claude-sonnet-5-5: 44k tokens per pass (3/3 passed)fix · task version 1 · codex · anthropic/claude-sonnet-5-544kfix · task version 1 · codex · baseten/glm-5.3-flash: 45k tokens per pass (3/3 passed)fix · task version 1 · codex · baseten/glm-5.3-flash45kfix · task version 1 · codex · fireworks/kimi-k3: 36k tokens per pass (3/3 passed)fix · task version 1 · codex · fireworks/kimi-k336kfix · task version 1 · codex · fireworks/qwen3.8-max: 56k tokens per pass (3/3 passed)fix · task version 1 · codex · fireworks/qwen3.8-max56kfix · task version 1 · codex · google/gemini-3.8-flash: 159k tokens per pass (3/3 passed)fix · task version 1 · codex · google/gemini-3.8-flash159kfix · task version 1 · codex · google/gemini-3.8-flash @high: 251k tokens per pass (3/3 passed)fix · task version 1 · codex · google/gemini-3.8-flash @high251kfix · task version 1 · codex · google/gemini-3.8-flash @low: 145k tokens per pass (3/3 passed)fix · task version 1 · codex · google/gemini-3.8-flash @low145kfix · task version 1 · codex · moonshot/kimi-k3: 46k tokens per pass (3/3 passed)fix · task version 1 · codex · moonshot/kimi-k346kfix · task version 1 · codex · together/deepseek-v4.1-flash: 60k tokens per pass (3/3 passed)fix · task version 1 · codex · together/deepseek-v4.1-flash60kfix · task version 1 · codex · openai/gpt-6.1-sol @high: 81k tokens per pass (3/3 passed)fix · task version 1 · codex · openai/gpt-6.1-sol @high81kfix · task version 1 · codex · openai/gpt-6.1-sol @medium: 72k tokens per pass (3/3 passed)fix · task version 1 · codex · openai/gpt-6.1-sol @medium72kfix · task version 1 · opencode · anthropic/claude-opus-5-5: 61k tokens per pass (3/3 passed)fix · task version 1 · opencode · anthropic/claude-opus-5-561kfix · task version 1 · opencode · anthropic/claude-sonnet-5-5: 61k tokens per pass (3/3 passed)fix · task version 1 · opencode · anthropic/claude-sonnet-5-561kfix · task version 1 · opencode · baseten/glm-5.3-flash: 69k tokens per pass (3/3 passed)fix · task version 1 · opencode · baseten/glm-5.3-flash69kfix · task version 1 · opencode · fireworks/kimi-k3: 66k tokens per pass (3/3 passed)fix · task version 1 · opencode · fireworks/kimi-k366kfix · task version 1 · opencode · fireworks/qwen3.8-max: 86k tokens per pass (3/3 passed)fix · task version 1 · opencode · fireworks/qwen3.8-max86kfix · task version 1 · opencode · google/gemini-3.8-flash: 217k tokens per pass (3/3 passed)fix · task version 1 · opencode · google/gemini-3.8-flash217kfix · task version 1 · opencode · google/gemini-3.8-flash @high: 446k tokens per pass (3/3 passed)fix · task version 1 · opencode · google/gemini-3.8-flash @high446kfix · task version 1 · opencode · google/gemini-3.8-flash @low: 135k tokens per pass (3/3 passed)fix · task version 1 · opencode · google/gemini-3.8-flash @low135kfix · task version 1 · opencode · moonshot/kimi-k3: 78k tokens per pass (3/3 passed)fix · task version 1 · opencode · moonshot/kimi-k378kfix · task version 1 · opencode · openai/gpt-6.1-sol: 108k tokens per pass (3/3 passed)fix · task version 1 · opencode · openai/gpt-6.1-sol108kfix · task version 1 · pi · anthropic/claude-opus-5-5: 16k tokens per pass (3/3 passed)fix · task version 1 · pi · anthropic/claude-opus-5-516kfix · task version 1 · pi · anthropic/claude-sonnet-5-5: 15k tokens per pass (3/3 passed)fix · task version 1 · pi · anthropic/claude-sonnet-5-515kfix · task version 1 · pi · baseten/glm-5.3-flash: 27k tokens per pass (3/3 passed)fix · task version 1 · pi · baseten/glm-5.3-flash27kfix · task version 1 · pi · fireworks/kimi-k3: 29k tokens per pass (3/3 passed)fix · task version 1 · pi · fireworks/kimi-k329kfix · task version 1 · pi · fireworks/qwen3.8-max: 55k tokens per pass (3/3 passed)fix · task version 1 · pi · fireworks/qwen3.8-max55kfix · task version 1 · pi · google/gemini-3.8-flash: 100k tokens per pass (3/3 passed)fix · task version 1 · pi · google/gemini-3.8-flash100kfix · task version 1 · pi · google/gemini-3.8-flash @high: 136k tokens per pass (3/3 passed)fix · task version 1 · pi · google/gemini-3.8-flash @high136kfix · task version 1 · pi · google/gemini-3.8-flash @low: 60k tokens per pass (3/3 passed)fix · task version 1 · pi · google/gemini-3.8-flash @low60kfix · task version 1 · pi · moonshot/kimi-k3: 27k tokens per pass (3/3 passed)fix · task version 1 · pi · moonshot/kimi-k327kfix · task version 1 · pi · openai/gpt-6.1-sol: 25k tokens per pass (3/3 passed)fix · task version 1 · pi · openai/gpt-6.1-sol25kfix · task version 1 · pi · together/deepseek-v4.1-flash: 23k tokens per pass (3/3 passed)fix · task version 1 · pi · together/deepseek-v4.1-flash23kfix · task version 1 · opencode · zai/glm-5.2: 129k tokens per pass (5/6 passed)fix · task version 1 · opencode · zai/glm-5.2129kfix · task version 1 · pi · zai/glm-5.2: 84k tokens per pass (5/6 passed)fix · task version 1 · pi · zai/glm-5.284kfix · task version 2 · codex · openai/gpt-6.1-sol @medium: 111k tokens per pass (13/13 passed)fix · task version 2 · codex · openai/gpt-6.1-sol @medium111kfix · task version 2 · codex · openai/gpt-6.1-sol @high: 122k tokens per pass (12/12 passed)fix · task version 2 · codex · openai/gpt-6.1-sol @high122kfix · task version 2 · pi · google/gemini-3.8-flash @high: 1.19M tokens per pass (6/6 passed)fix · task version 2 · pi · google/gemini-3.8-flash @high1.19Mfix · task version 2 · codex · fireworks/kimi-k3: 58k tokens per pass (5/5 passed)fix · task version 2 · codex · fireworks/kimi-k358kfix · task version 2 · codex · openai/gpt-6.1-sol: 106k tokens per pass (5/5 passed)fix · task version 2 · codex · openai/gpt-6.1-sol106kfix · task version 2 · codex · openai/gpt-6.1-sol @xhigh: 145k tokens per pass (5/5 passed)fix · task version 2 · codex · openai/gpt-6.1-sol @xhigh145kfix · task version 2 · opencode · baseten/glm-5.3-flash: 110k tokens per pass (5/5 passed)fix · task version 2 · opencode · baseten/glm-5.3-flash110kfix · task version 2 · opencode · openai/gpt-6.1-sol: 130k tokens per pass (5/5 passed)fix · task version 2 · opencode · openai/gpt-6.1-sol130kfix · task version 2 · pi · baseten/glm-5.3-flash: 53k tokens per pass (5/5 passed)fix · task version 2 · pi · baseten/glm-5.3-flash53kfix · task version 2 · pi · fireworks/qwen3.8-max: 118k tokens per pass (5/5 passed)fix · task version 2 · pi · fireworks/qwen3.8-max118kfix · task version 2 · pi · google/gemini-3.8-flash: 879k tokens per pass (5/5 passed)fix · task version 2 · pi · google/gemini-3.8-flash879kfix · task version 2 · claude-code · fireworks/qwen3.8-max: 298k tokens per pass (4/4 passed)fix · task version 2 · claude-code · fireworks/qwen3.8-max298kfix · task version 2 · claude-code · together/deepseek-v4.1-flash: 374k tokens per pass (4/4 passed)fix · task version 2 · claude-code · together/deepseek-v4.1-flash374kfix · task version 2 · antigravity · google/gemini-3.1-pro @high: 193k tokens per pass (3/3 passed)fix · task version 2 · antigravity · google/gemini-3.1-pro @high193kfix · task version 2 · antigravity · google/gemini-3.8-flash @high: 1.17M tokens per pass (3/3 passed)fix · task version 2 · antigravity · google/gemini-3.8-flash @high1.17Mfix · task version 2 · antigravity · google/gemini-4-argon @high: 187k tokens per pass (3/3 passed)fix · task version 2 · antigravity · google/gemini-4-argon @high187kfix · task version 2 · claude-code · anthropic/claude-sonnet-5-5: 765k tokens per pass (3/3 passed)fix · task version 2 · claude-code · anthropic/claude-sonnet-5-5765kfix · task version 2 · claude-code · baseten/glm-5.3-flash: 373k tokens per pass (3/3 passed)fix · task version 2 · claude-code · baseten/glm-5.3-flash373kfix · task version 2 · claude-code · google/gemini-3.8-flash @high: 1.33M tokens per pass (3/3 passed)fix · task version 2 · claude-code · google/gemini-3.8-flash @high1.33Mfix · task version 2 · claude-code · google/gemini-3.8-flash @medium: 514k tokens per pass (3/3 passed)fix · task version 2 · claude-code · google/gemini-3.8-flash @medium514kfix · task version 2 · claude-code · openai/gpt-6.1-sol: 191k tokens per pass (3/3 passed)fix · task version 2 · claude-code · openai/gpt-6.1-sol191kfix · task version 2 · codex · anthropic/claude-sonnet-5-5: 46k tokens per pass (3/3 passed)fix · task version 2 · codex · anthropic/claude-sonnet-5-546kfix · task version 2 · codex · fireworks/qwen3.8-max: 106k tokens per pass (3/3 passed)fix · task version 2 · codex · fireworks/qwen3.8-max106kfix · task version 2 · codex · google/gemini-3.8-flash @high: 809k tokens per pass (3/3 passed)fix · task version 2 · codex · google/gemini-3.8-flash @high809kfix · task version 2 · codex · google/gemini-3.8-flash @low: 319k tokens per pass (3/3 passed)fix · task version 2 · codex · google/gemini-3.8-flash @low319kfix · task version 2 · opencode · anthropic/claude-sonnet-5-5: 55k tokens per pass (3/3 passed)fix · task version 2 · opencode · anthropic/claude-sonnet-5-555kfix · task version 2 · opencode · fireworks/kimi-k3: 88k tokens per pass (3/3 passed)fix · task version 2 · opencode · fireworks/kimi-k388kfix · task version 2 · opencode · fireworks/qwen3.8-max: 123k tokens per pass (3/3 passed)fix · task version 2 · opencode · fireworks/qwen3.8-max123kfix · task version 2 · opencode · google/gemini-3.8-flash: 645k tokens per pass (3/3 passed)fix · task version 2 · opencode · google/gemini-3.8-flash645kfix · task version 2 · opencode · google/gemini-3.8-flash @high: 943k tokens per pass (3/3 passed)fix · task version 2 · opencode · google/gemini-3.8-flash @high943kfix · task version 2 · opencode · google/gemini-3.8-flash @low: 225k tokens per pass (3/3 passed)fix · task version 2 · opencode · google/gemini-3.8-flash @low225kfix · task version 2 · opencode · moonshot/kimi-k3: 94k tokens per pass (3/3 passed)fix · task version 2 · opencode · moonshot/kimi-k394kfix · task version 2 · opencode · zai/glm-5.2: 123k tokens per pass (3/3 passed)fix · task version 2 · opencode · zai/glm-5.2123kfix · task version 2 · pi · anthropic/claude-sonnet-5-5: 19k tokens per pass (3/3 passed)fix · task version 2 · pi · anthropic/claude-sonnet-5-519kfix · task version 2 · pi · fireworks/kimi-k3: 37k tokens per pass (3/3 passed)fix · task version 2 · pi · fireworks/kimi-k337kfix · task version 2 · pi · google/gemini-3.8-flash @low: 105k tokens per pass (3/3 passed)fix · task version 2 · pi · google/gemini-3.8-flash @low105kfix · task version 2 · pi · moonshot/kimi-k3: 39k tokens per pass (3/3 passed)fix · task version 2 · pi · moonshot/kimi-k339kfix · task version 2 · pi · openai/gpt-6.1-sol: 34k tokens per pass (3/3 passed)fix · task version 2 · pi · openai/gpt-6.1-sol34kfix · task version 2 · claude-code · moonshot/kimi-k3: 215k tokens per pass (2/2 passed)fix · task version 2 · claude-code · moonshot/kimi-k3215kfix · task version 2 · pi · together/deepseek-v4.1-flash: 57k tokens per pass (5/6 passed)fix · task version 2 · pi · together/deepseek-v4.1-flash57kfix · task version 2 · claude-code · anthropic/claude-opus-5-5: 2.18M tokens per pass (4/6 passed)fix · task version 2 · claude-code · anthropic/claude-opus-5-52.18Mfix · task version 2 · claude-code · fireworks/kimi-k3: 335k tokens per pass (2/3 passed)fix · task version 2 · claude-code · fireworks/kimi-k3335kfix · task version 2 · claude-code · google/gemini-3.8-flash: 1.05M tokens per pass (2/3 passed)fix · task version 2 · claude-code · google/gemini-3.8-flash1.05Mfix · task version 2 · claude-code · google/gemini-3.8-flash @low: 318k tokens per pass (2/3 passed)fix · task version 2 · claude-code · google/gemini-3.8-flash @low318kfix · task version 2 · claude-code · zai/glm-5.2: 758k tokens per pass (2/3 passed)fix · task version 2 · claude-code · zai/glm-5.2758kfix · task version 2 · codex · anthropic/claude-opus-5-5: 67k tokens per pass (2/3 passed)fix · task version 2 · codex · anthropic/claude-opus-5-567kfix · task version 2 · codex · baseten/glm-5.3-flash: 91k tokens per pass (2/3 passed)fix · task version 2 · codex · baseten/glm-5.3-flash91kfix · task version 2 · codex · google/gemini-3.8-flash: 246k tokens per pass (2/3 passed)fix · task version 2 · codex · google/gemini-3.8-flash246kfix · task version 2 · codex · together/deepseek-v4.1-flash: 90k tokens per pass (2/3 passed)fix · task version 2 · codex · together/deepseek-v4.1-flash90kfix · task version 2 · opencode · anthropic/claude-opus-5-5: 78k tokens per pass (2/3 passed)fix · task version 2 · opencode · anthropic/claude-opus-5-578kfix · task version 2 · opencode · together/deepseek-v4.1-flash: 108k tokens per pass (2/3 passed)fix · task version 2 · opencode · together/deepseek-v4.1-flash108kfix · task version 2 · pi · anthropic/claude-opus-5-5: 31k tokens per pass (2/3 passed)fix · task version 2 · pi · anthropic/claude-opus-5-531kfix · task version 2 · pi · zai/glm-5.2: 99k tokens per pass (2/3 passed)fix · task version 2 · pi · zai/glm-5.299kfix · task version 2 · codex · zai/glm-5.2: 243k tokens per pass (2/4 passed)fix · task version 2 · codex · zai/glm-5.2243kfix · task version 2 · claude-code · anthropic/claude-opus-5-5 @xhigh: 4.67M tokens per pass (2/5 passed)fix · task version 2 · claude-code · anthropic/claude-opus-5-5 @xhigh4.67Mfix · task version 2 · codex · moonshot/kimi-k3: 212k tokens per pass (1/3 passed)fix · task version 2 · codex · moonshot/kimi-k3212khard-fix · codex · openai/gpt-6.1-sol @medium: 97k tokens per pass (26/26 passed)hard-fix · codex · openai/gpt-6.1-sol @medium97khard-fix · codex · openai/gpt-6.1-sol @high: 125k tokens per pass (20/20 passed)hard-fix · codex · openai/gpt-6.1-sol @high125khard-fix · claude-code · anthropic/claude-opus-5-5: 1.06M tokens per pass (15/15 passed)hard-fix · claude-code · anthropic/claude-opus-5-51.06Mhard-fix · opencode · openai/gpt-6.1-sol: 129k tokens per pass (13/13 passed)hard-fix · opencode · openai/gpt-6.1-sol129khard-fix · pi · fireworks/qwen3.8-max: 96k tokens per pass (13/13 passed)hard-fix · pi · fireworks/qwen3.8-max96khard-fix · pi · google/gemini-3.8-flash: 456k tokens per pass (13/13 passed)hard-fix · pi · google/gemini-3.8-flash456khard-fix · claude-code · anthropic/claude-sonnet-5-5: 804k tokens per pass (11/11 passed)hard-fix · claude-code · anthropic/claude-sonnet-5-5804khard-fix · claude-code · openai/gpt-6.1-sol: 223k tokens per pass (11/11 passed)hard-fix · claude-code · openai/gpt-6.1-sol223khard-fix · codex · anthropic/claude-opus-5-5: 57k tokens per pass (11/11 passed)hard-fix · codex · anthropic/claude-opus-5-557khard-fix · codex · anthropic/claude-sonnet-5-5: 42k tokens per pass (11/11 passed)hard-fix · codex · anthropic/claude-sonnet-5-542khard-fix · codex · baseten/glm-5.3-flash: 62k tokens per pass (11/11 passed)hard-fix · codex · baseten/glm-5.3-flash62khard-fix · codex · fireworks/qwen3.8-max: 133k tokens per pass (11/11 passed)hard-fix · codex · fireworks/qwen3.8-max133khard-fix · codex · google/gemini-3.8-flash: 422k tokens per pass (11/11 passed)hard-fix · codex · google/gemini-3.8-flash422khard-fix · codex · google/gemini-3.8-flash @high: 828k tokens per pass (11/11 passed)hard-fix · codex · google/gemini-3.8-flash @high828khard-fix · opencode · anthropic/claude-opus-5-5: 58k tokens per pass (11/11 passed)hard-fix · opencode · anthropic/claude-opus-5-558khard-fix · opencode · google/gemini-3.8-flash: 508k tokens per pass (11/11 passed)hard-fix · opencode · google/gemini-3.8-flash508khard-fix · pi · anthropic/claude-opus-5-5: 24k tokens per pass (11/11 passed)hard-fix · pi · anthropic/claude-opus-5-524khard-fix · pi · anthropic/claude-sonnet-5-5: 18k tokens per pass (11/11 passed)hard-fix · pi · anthropic/claude-sonnet-5-518khard-fix · pi · openai/gpt-6.1-sol: 38k tokens per pass (11/11 passed)hard-fix · pi · openai/gpt-6.1-sol38khard-fix · claude-code · anthropic/claude-opus-5-5 @xhigh: 2.06M tokens per pass (5/5 passed)hard-fix · claude-code · anthropic/claude-opus-5-5 @xhigh2.06Mhard-fix · codex · openai/gpt-6.1-sol @xhigh: 142k tokens per pass (5/5 passed)hard-fix · codex · openai/gpt-6.1-sol @xhigh142khard-fix · antigravity · google/gemini-4-argon @high: 238k tokens per pass (15/16 passed)hard-fix · antigravity · google/gemini-4-argon @high238khard-fix · claude-code · google/gemini-3.8-flash @low: 312k tokens per pass (15/16 passed)hard-fix · claude-code · google/gemini-3.8-flash @low312khard-fix · claude-code · together/deepseek-v4.1-flash: 429k tokens per pass (12/13 passed)hard-fix · claude-code · together/deepseek-v4.1-flash429khard-fix · pi · google/gemini-3.8-flash @high: 947k tokens per pass (12/13 passed)hard-fix · pi · google/gemini-3.8-flash @high947khard-fix · antigravity · google/gemini-3.8-flash @high: 1.10M tokens per pass (10/11 passed)hard-fix · antigravity · google/gemini-3.8-flash @high1.10Mhard-fix · claude-code · baseten/glm-5.3-flash: 432k tokens per pass (10/11 passed)hard-fix · claude-code · baseten/glm-5.3-flash432khard-fix · claude-code · fireworks/kimi-k3: 306k tokens per pass (10/11 passed)hard-fix · claude-code · fireworks/kimi-k3306khard-fix · claude-code · google/gemini-3.8-flash: 1.32M tokens per pass (10/11 passed)hard-fix · claude-code · google/gemini-3.8-flash1.32Mhard-fix · claude-code · moonshot/kimi-k3: 457k tokens per pass (10/11 passed)hard-fix · claude-code · moonshot/kimi-k3457khard-fix · codex · moonshot/kimi-k3: 123k tokens per pass (10/11 passed)hard-fix · codex · moonshot/kimi-k3123khard-fix · codex · together/deepseek-v4.1-flash: 95k tokens per pass (10/11 passed)hard-fix · codex · together/deepseek-v4.1-flash95khard-fix · opencode · anthropic/claude-sonnet-5-5: 68k tokens per pass (10/11 passed)hard-fix · opencode · anthropic/claude-sonnet-5-568khard-fix · opencode · fireworks/qwen3.8-max: 238k tokens per pass (10/11 passed)hard-fix · opencode · fireworks/qwen3.8-max238khard-fix · opencode · zai/glm-5.2: 240k tokens per pass (10/11 passed)hard-fix · opencode · zai/glm-5.2240khard-fix · pi · fireworks/kimi-k3: 54k tokens per pass (10/11 passed)hard-fix · pi · fireworks/kimi-k354khard-fix · pi · moonshot/kimi-k3: 62k tokens per pass (10/11 passed)hard-fix · pi · moonshot/kimi-k362khard-fix · pi · zai/glm-5.2: 134k tokens per pass (10/11 passed)hard-fix · pi · zai/glm-5.2134khard-fix · claude-code · google/gemini-3.8-flash @medium: 787k tokens per pass (14/16 passed)hard-fix · claude-code · google/gemini-3.8-flash @medium787khard-fix · opencode · google/gemini-3.8-flash @high: 1.30M tokens per pass (14/16 passed)hard-fix · opencode · google/gemini-3.8-flash @high1.30Mhard-fix · opencode · google/gemini-3.8-flash @low: 248k tokens per pass (14/16 passed)hard-fix · opencode · google/gemini-3.8-flash @low248khard-fix · pi · google/gemini-3.8-flash @low: 110k tokens per pass (14/16 passed)hard-fix · pi · google/gemini-3.8-flash @low110khard-fix · opencode · moonshot/kimi-k3: 151k tokens per pass (9/11 passed)hard-fix · opencode · moonshot/kimi-k3151khard-fix · claude-code · google/gemini-3.8-flash @high: 1.79M tokens per pass (13/16 passed)hard-fix · claude-code · google/gemini-3.8-flash @high1.79Mhard-fix · codex · google/gemini-3.8-flash @low: 333k tokens per pass (13/16 passed)hard-fix · codex · google/gemini-3.8-flash @low333khard-fix · pi · together/deepseek-v4.1-flash: 65k tokens per pass (12/15 passed)hard-fix · pi · together/deepseek-v4.1-flash65khard-fix · codex · fireworks/kimi-k3: 89k tokens per pass (10/13 passed)hard-fix · codex · fireworks/kimi-k389khard-fix · codex · zai/glm-5.2: 159k tokens per pass (10/13 passed)hard-fix · codex · zai/glm-5.2159khard-fix · opencode · fireworks/kimi-k3: 176k tokens per pass (8/11 passed)hard-fix · opencode · fireworks/kimi-k3176khard-fix · opencode · baseten/glm-5.3-flash: 162k tokens per pass (9/13 passed)hard-fix · opencode · baseten/glm-5.3-flash162khard-fix · pi · baseten/glm-5.3-flash: 61k tokens per pass (9/13 passed)hard-fix · pi · baseten/glm-5.3-flash61khard-review · task version 1 · pi · together/deepseek-v4.1-flash: 244k tokens per pass (13/13 passed)hard-review · task version 1 · pi · together/deepseek-v4.1-flash244khard-review · task version 1 · codex · zai/glm-5.2: 490k tokens per pass (16/18 passed)hard-review · task version 1 · codex · zai/glm-5.2490khard-review · task version 1 · antigravity · google/gemini-3.8-flash @high: 1.77M tokens per pass (21/24 passed)hard-review · task version 1 · antigravity · google/gemini-3.8-flash @high1.77Mhard-review · task version 1 · antigravity · google/gemini-4-argon @high: 272k tokens per pass (7/8 passed)hard-review · task version 1 · antigravity · google/gemini-4-argon @high272khard-review · task version 1 · opencode · fireworks/kimi-k3: 231k tokens per pass (7/8 passed)hard-review · task version 1 · opencode · fireworks/kimi-k3231khard-review · task version 1 · opencode · zai/glm-5.2: 294k tokens per pass (7/8 passed)hard-review · task version 1 · opencode · zai/glm-5.2294khard-review · task version 1 · pi · fireworks/kimi-k3: 159k tokens per pass (7/8 passed)hard-review · task version 1 · pi · fireworks/kimi-k3159khard-review · task version 1 · pi · google/gemini-3.8-flash: 1.08M tokens per pass (7/8 passed)hard-review · task version 1 · pi · google/gemini-3.8-flash1.08Mhard-review · task version 1 · pi · google/gemini-3.8-flash @high: 1.89M tokens per pass (7/8 passed)hard-review · task version 1 · pi · google/gemini-3.8-flash @high1.89Mhard-review · task version 1 · pi · moonshot/kimi-k3: 147k tokens per pass (7/8 passed)hard-review · task version 1 · pi · moonshot/kimi-k3147khard-review · task version 1 · pi · openai/gpt-6.1-sol: 77k tokens per pass (7/8 passed)hard-review · task version 1 · pi · openai/gpt-6.1-sol77khard-review · task version 1 · pi · zai/glm-5.2: 305k tokens per pass (7/8 passed)hard-review · task version 1 · pi · zai/glm-5.2305khard-review · task version 1 · claude-code · together/deepseek-v4.1-flash: 816k tokens per pass (15/18 passed)hard-review · task version 1 · claude-code · together/deepseek-v4.1-flash816khard-review · task version 1 · opencode · baseten/glm-5.3-flash: 251k tokens per pass (15/18 passed)hard-review · task version 1 · opencode · baseten/glm-5.3-flash251khard-review · task version 1 · pi · baseten/glm-5.3-flash: 203k tokens per pass (14/18 passed)hard-review · task version 1 · pi · baseten/glm-5.3-flash203khard-review · task version 1 · pi · fireworks/qwen3.8-max: 377k tokens per pass (14/18 passed)hard-review · task version 1 · pi · fireworks/qwen3.8-max377khard-review · task version 1 · codex · openai/gpt-6.1-sol @medium: 185k tokens per pass (38/50 passed)hard-review · task version 1 · codex · openai/gpt-6.1-sol @medium185khard-review · task version 1 · claude-code · baseten/glm-5.3-flash: 1.27M tokens per pass (6/8 passed)hard-review · task version 1 · claude-code · baseten/glm-5.3-flash1.27Mhard-review · task version 1 · claude-code · fireworks/kimi-k3: 327k tokens per pass (6/8 passed)hard-review · task version 1 · claude-code · fireworks/kimi-k3327khard-review · task version 1 · claude-code · openai/gpt-6.1-sol: 390k tokens per pass (6/8 passed)hard-review · task version 1 · claude-code · openai/gpt-6.1-sol390khard-review · task version 1 · codex · google/gemini-3.8-flash: 1.28M tokens per pass (6/8 passed)hard-review · task version 1 · codex · google/gemini-3.8-flash1.28Mhard-review · task version 1 · opencode · anthropic/claude-sonnet-5-5: 164k tokens per pass (6/8 passed)hard-review · task version 1 · opencode · anthropic/claude-sonnet-5-5164khard-review · task version 1 · opencode · fireworks/qwen3.8-max: 396k tokens per pass (6/8 passed)hard-review · task version 1 · opencode · fireworks/qwen3.8-max396khard-review · task version 1 · opencode · moonshot/kimi-k3: 244k tokens per pass (6/8 passed)hard-review · task version 1 · opencode · moonshot/kimi-k3244khard-review · task version 1 · claude-code · google/gemini-3.8-flash: 2.40M tokens per pass (5/8 passed)hard-review · task version 1 · claude-code · google/gemini-3.8-flash2.40Mhard-review · task version 1 · claude-code · moonshot/kimi-k3: 599k tokens per pass (5/8 passed)hard-review · task version 1 · claude-code · moonshot/kimi-k3599khard-review · task version 1 · codex · anthropic/claude-opus-5-5: 145k tokens per pass (5/8 passed)hard-review · task version 1 · codex · anthropic/claude-opus-5-5145khard-review · task version 1 · codex · moonshot/kimi-k3: 248k tokens per pass (5/8 passed)hard-review · task version 1 · codex · moonshot/kimi-k3248khard-review · task version 1 · opencode · google/gemini-3.8-flash: 1.45M tokens per pass (5/8 passed)hard-review · task version 1 · opencode · google/gemini-3.8-flash1.45Mhard-review · task version 1 · pi · anthropic/claude-sonnet-5-5: 98k tokens per pass (5/8 passed)hard-review · task version 1 · pi · anthropic/claude-sonnet-5-598khard-review · task version 1 · codex · fireworks/kimi-k3: 210k tokens per pass (11/18 passed)hard-review · task version 1 · codex · fireworks/kimi-k3210khard-review · task version 1 · codex · openai/gpt-6.1-sol @xhigh: 336k tokens per pass (14/25 passed)hard-review · task version 1 · codex · openai/gpt-6.1-sol @xhigh336khard-review · task version 1 · opencode · openai/gpt-6.1-sol: 437k tokens per pass (9/18 passed)hard-review · task version 1 · opencode · openai/gpt-6.1-sol437khard-review · task version 1 · codex · anthropic/claude-sonnet-5-5: 168k tokens per pass (4/8 passed)hard-review · task version 1 · codex · anthropic/claude-sonnet-5-5168khard-review · task version 1 · codex · fireworks/qwen3.8-max: 594k tokens per pass (4/8 passed)hard-review · task version 1 · codex · fireworks/qwen3.8-max594khard-review · task version 1 · pi · anthropic/claude-opus-5-5: 117k tokens per pass (4/8 passed)hard-review · task version 1 · pi · anthropic/claude-opus-5-5117khard-review · task version 1 · codex · openai/gpt-6.1-sol @high: 320k tokens per pass (17/35 passed)hard-review · task version 1 · codex · openai/gpt-6.1-sol @high320khard-review · task version 1 · claude-code · anthropic/claude-opus-5-5 @xhigh: 4.53M tokens per pass (8/20 passed)hard-review · task version 1 · claude-code · anthropic/claude-opus-5-5 @xhigh4.53Mhard-review · task version 1 · claude-code · anthropic/claude-opus-5-5: 3.00M tokens per pass (11/28 passed)hard-review · task version 1 · claude-code · anthropic/claude-opus-5-53.00Mhard-review · task version 1 · claude-code · anthropic/claude-sonnet-5-5: 2.53M tokens per pass (3/8 passed)hard-review · task version 1 · claude-code · anthropic/claude-sonnet-5-52.53Mhard-review · task version 1 · codex · baseten/glm-5.3-flash: 485k tokens per pass (3/8 passed)hard-review · task version 1 · codex · baseten/glm-5.3-flash485khard-review · task version 1 · opencode · anthropic/claude-opus-5-5: 262k tokens per pass (3/8 passed)hard-review · task version 1 · opencode · anthropic/claude-opus-5-5262khard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @max: 892k tokens per pass (4/4 passed)hard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @max892khard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5: 788k tokens per pass (2/2 passed)hard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5788khard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @max: 1.02M tokens per pass (1/1 passed)hard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @max1.02Mhard-review · task version 2 · pi · anthropic/claude-opus-5-5 @max: 427k tokens per pass (1/1 passed)hard-review · task version 2 · pi · anthropic/claude-opus-5-5 @max427khard-review · task version 2 · pi · together/deepseek-v4.1-flash: 216k tokens per pass (31/32 passed)hard-review · task version 2 · pi · together/deepseek-v4.1-flash216khard-review · task version 2 · codex · zai/glm-5.2: 482k tokens per pass (20/22 passed)hard-review · task version 2 · codex · zai/glm-5.2482khard-review · task version 2 · opencode · together/deepseek-v4.1-flash: 142k tokens per pass (10/11 passed)hard-review · task version 2 · opencode · together/deepseek-v4.1-flash142khard-review · task version 2 · claude-code · zai/glm-5.2: 1.37M tokens per pass (9/10 passed)hard-review · task version 2 · claude-code · zai/glm-5.21.37Mhard-review · task version 2 · codex · together/deepseek-v4.1-flash: 216k tokens per pass (9/10 passed)hard-review · task version 2 · codex · together/deepseek-v4.1-flash216khard-review · task version 2 · opencode · fireworks/kimi-k3: 213k tokens per pass (9/10 passed)hard-review · task version 2 · opencode · fireworks/kimi-k3213khard-review · task version 2 · opencode · zai/glm-5.2: 268k tokens per pass (9/10 passed)hard-review · task version 2 · opencode · zai/glm-5.2268khard-review · task version 2 · pi · fireworks/kimi-k3: 142k tokens per pass (9/10 passed)hard-review · task version 2 · pi · fireworks/kimi-k3142khard-review · task version 2 · pi · moonshot/kimi-k3: 148k tokens per pass (9/10 passed)hard-review · task version 2 · pi · moonshot/kimi-k3148khard-review · task version 2 · antigravity · google/gemini-3.8-flash @high: 1.67M tokens per pass (24/27 passed)hard-review · task version 2 · antigravity · google/gemini-3.8-flash @high1.67Mhard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @max: 6.10M tokens per pass (8/9 passed)hard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @max6.10Mhard-review · task version 2 · codex · anthropic/claude-opus-5-5 @high: 130k tokens per pass (8/9 passed)hard-review · task version 2 · codex · anthropic/claude-opus-5-5 @high130khard-review · task version 2 · antigravity · google/gemini-3.8-flash @medium: 1.50M tokens per pass (7/8 passed)hard-review · task version 2 · antigravity · google/gemini-3.8-flash @medium1.50Mhard-review · task version 2 · antigravity · google/gemini-4-argon @medium: 144k tokens per pass (7/8 passed)hard-review · task version 2 · antigravity · google/gemini-4-argon @medium144khard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @max: 12.51M tokens per pass (7/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @max12.51Mhard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @xhigh: 1.91M tokens per pass (7/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @xhigh1.91Mhard-review · task version 2 · claude-code · google/gemini-3.8-flash @medium: 938k tokens per pass (7/8 passed)hard-review · task version 2 · claude-code · google/gemini-3.8-flash @medium938khard-review · task version 2 · claude-code · openai/gpt-5.6-terra @low: 197k tokens per pass (7/8 passed)hard-review · task version 2 · claude-code · openai/gpt-5.6-terra @low197khard-review · task version 2 · claude-code · openai/gpt-6-luna @high: 332k tokens per pass (7/8 passed)hard-review · task version 2 · claude-code · openai/gpt-6-luna @high332khard-review · task version 2 · claude-code · openai/gpt-6.1-sol @low: 263k tokens per pass (7/8 passed)hard-review · task version 2 · claude-code · openai/gpt-6.1-sol @low263khard-review · task version 2 · codex · google/gemini-3.8-flash @high: 1.72M tokens per pass (7/8 passed)hard-review · task version 2 · codex · google/gemini-3.8-flash @high1.72Mhard-review · task version 2 · codex · openai/gpt-6-luna @xhigh: 242k tokens per pass (7/8 passed)hard-review · task version 2 · codex · openai/gpt-6-luna @xhigh242khard-review · task version 2 · codex · openai/gpt-6.1-sol @low: 157k tokens per pass (7/8 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol @low157khard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @xhigh: 378k tokens per pass (7/8 passed)hard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @xhigh378khard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @xhigh: 279k tokens per pass (7/8 passed)hard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @xhigh279khard-review · task version 2 · opencode · google/gemini-3.8-flash @high: 1.88M tokens per pass (7/8 passed)hard-review · task version 2 · opencode · google/gemini-3.8-flash @high1.88Mhard-review · task version 2 · opencode · google/gemini-3.8-flash @medium: 1.12M tokens per pass (7/8 passed)hard-review · task version 2 · opencode · google/gemini-3.8-flash @medium1.12Mhard-review · task version 2 · opencode · openai/gpt-6-luna @low: 97k tokens per pass (7/8 passed)hard-review · task version 2 · opencode · openai/gpt-6-luna @low97khard-review · task version 2 · pi · anthropic/claude-opus-5-5 @xhigh: 157k tokens per pass (7/8 passed)hard-review · task version 2 · pi · anthropic/claude-opus-5-5 @xhigh157khard-review · task version 2 · claude-code · together/deepseek-v4.1-flash: 705k tokens per pass (19/22 passed)hard-review · task version 2 · claude-code · together/deepseek-v4.1-flash705khard-review · task version 2 · opencode · baseten/glm-5.3-flash: 235k tokens per pass (19/22 passed)hard-review · task version 2 · opencode · baseten/glm-5.3-flash235khard-review · task version 2 · codex · anthropic/claude-opus-5-5 @max: 886k tokens per pass (6/7 passed)hard-review · task version 2 · codex · anthropic/claude-opus-5-5 @max886khard-review · task version 2 · pi · google/gemini-3.8-flash: 1.12M tokens per pass (17/20 passed)hard-review · task version 2 · pi · google/gemini-3.8-flash1.12Mhard-review · task version 2 · pi · google/gemini-3.8-flash @high: 1.67M tokens per pass (10/12 passed)hard-review · task version 2 · pi · google/gemini-3.8-flash @high1.67Mhard-review · task version 2 · codex · openai/gpt-6.1-sol @medium: 170k tokens per pass (56/68 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol @medium170khard-review · task version 2 · pi · baseten/glm-5.3-flash: 189k tokens per pass (18/22 passed)hard-review · task version 2 · pi · baseten/glm-5.3-flash189khard-review · task version 2 · pi · fireworks/qwen3.8-max: 331k tokens per pass (17/21 passed)hard-review · task version 2 · pi · fireworks/qwen3.8-max331khard-review · task version 2 · claude-code · fireworks/kimi-k3: 290k tokens per pass (8/10 passed)hard-review · task version 2 · claude-code · fireworks/kimi-k3290khard-review · task version 2 · codex · google/gemini-3.8-flash: 1.11M tokens per pass (8/10 passed)hard-review · task version 2 · codex · google/gemini-3.8-flash1.11Mhard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5: 151k tokens per pass (8/10 passed)hard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5151khard-review · task version 2 · opencode · moonshot/kimi-k3: 229k tokens per pass (8/10 passed)hard-review · task version 2 · opencode · moonshot/kimi-k3229khard-review · task version 2 · pi · openai/gpt-6.1-sol: 86k tokens per pass (8/10 passed)hard-review · task version 2 · pi · openai/gpt-6.1-sol86khard-review · task version 2 · pi · zai/glm-5.2: 325k tokens per pass (8/10 passed)hard-review · task version 2 · pi · zai/glm-5.2325khard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @xhigh: 2.26M tokens per pass (22/28 passed)hard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @xhigh2.26Mhard-review · task version 2 · antigravity · google/gemini-4-argon @high: 310k tokens per pass (7/9 passed)hard-review · task version 2 · antigravity · google/gemini-4-argon @high310khard-review · task version 2 · codex · anthropic/claude-opus-5-5 @xhigh: 269k tokens per pass (7/9 passed)hard-review · task version 2 · codex · anthropic/claude-opus-5-5 @xhigh269khard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @xhigh: 227k tokens per pass (7/9 passed)hard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @xhigh227khard-review · task version 2 · codex · fireworks/qwen3.8-max: 369k tokens per pass (7/9 passed)hard-review · task version 2 · codex · fireworks/qwen3.8-max369khard-review · task version 2 · opencode · fireworks/qwen3.8-max: 359k tokens per pass (7/9 passed)hard-review · task version 2 · opencode · fireworks/qwen3.8-max359khard-review · task version 2 · opencode · openai/gpt-6-luna @xhigh: 441k tokens per pass (7/9 passed)hard-review · task version 2 · opencode · openai/gpt-6-luna @xhigh441khard-review · task version 2 · opencode · openai/gpt-6.1-sol @low: 121k tokens per pass (7/9 passed)hard-review · task version 2 · opencode · openai/gpt-6.1-sol @low121khard-review · task version 2 · pi · openai/gpt-6.1-sol @low: 63k tokens per pass (7/9 passed)hard-review · task version 2 · pi · openai/gpt-6.1-sol @low63khard-review · task version 2 · antigravity · google/gemini-4-argon @low: 153k tokens per pass (6/8 passed)hard-review · task version 2 · antigravity · google/gemini-4-argon @low153khard-review · task version 2 · claude-code · openai/gpt-5.6-terra @high: 296k tokens per pass (6/8 passed)hard-review · task version 2 · claude-code · openai/gpt-5.6-terra @high296khard-review · task version 2 · codex · google/gemini-3.8-flash @medium: 1.26M tokens per pass (6/8 passed)hard-review · task version 2 · codex · google/gemini-3.8-flash @medium1.26Mhard-review · task version 2 · codex · openai/gpt-5.6-terra @high: 343k tokens per pass (6/8 passed)hard-review · task version 2 · codex · openai/gpt-5.6-terra @high343khard-review · task version 2 · codex · openai/gpt-5.6-terra @medium: 258k tokens per pass (6/8 passed)hard-review · task version 2 · codex · openai/gpt-5.6-terra @medium258khard-review · task version 2 · codex · openai/gpt-5.6-terra @xhigh: 366k tokens per pass (6/8 passed)hard-review · task version 2 · codex · openai/gpt-5.6-terra @xhigh366khard-review · task version 2 · codex · openai/gpt-6-luna @high: 226k tokens per pass (6/8 passed)hard-review · task version 2 · codex · openai/gpt-6-luna @high226khard-review · task version 2 · codex · openai/gpt-6-luna @max: 264k tokens per pass (6/8 passed)hard-review · task version 2 · codex · openai/gpt-6-luna @max264khard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @medium: 183k tokens per pass (6/8 passed)hard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @medium183khard-review · task version 2 · opencode · openai/gpt-5.6-terra @low: 194k tokens per pass (6/8 passed)hard-review · task version 2 · opencode · openai/gpt-5.6-terra @low194khard-review · task version 2 · opencode · openai/gpt-6-luna @high: 282k tokens per pass (6/8 passed)hard-review · task version 2 · opencode · openai/gpt-6-luna @high282khard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @high: 151k tokens per pass (6/8 passed)hard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @high151khard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @high: 70k tokens per pass (6/8 passed)hard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @high70khard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @xhigh: 93k tokens per pass (6/8 passed)hard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @xhigh93khard-review · task version 2 · pi · google/gemini-3.8-flash @medium: 1.35M tokens per pass (6/8 passed)hard-review · task version 2 · pi · google/gemini-3.8-flash @medium1.35Mhard-review · task version 2 · pi · openai/gpt-6-luna @high: 106k tokens per pass (6/8 passed)hard-review · task version 2 · pi · openai/gpt-6-luna @high106khard-review · task version 2 · codex · openai/gpt-6.1-sol: 196k tokens per pass (23/32 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol196khard-review · task version 2 · pi · google/gemini-3.8-flash @low: 434k tokens per pass (5/7 passed)hard-review · task version 2 · pi · google/gemini-3.8-flash @low434khard-review · task version 2 · claude-code · fireworks/qwen3.8-max: 430k tokens per pass (17/24 passed)hard-review · task version 2 · claude-code · fireworks/qwen3.8-max430khard-review · task version 2 · claude-code · baseten/glm-5.3-flash: 1.19M tokens per pass (7/10 passed)hard-review · task version 2 · claude-code · baseten/glm-5.3-flash1.19Mhard-review · task version 2 · claude-code · moonshot/kimi-k3: 485k tokens per pass (7/10 passed)hard-review · task version 2 · claude-code · moonshot/kimi-k3485khard-review · task version 2 · claude-code · openai/gpt-6.1-sol: 387k tokens per pass (7/10 passed)hard-review · task version 2 · claude-code · openai/gpt-6.1-sol387khard-review · task version 2 · opencode · google/gemini-3.8-flash: 1.30M tokens per pass (7/10 passed)hard-review · task version 2 · opencode · google/gemini-3.8-flash1.30Mhard-review · task version 2 · codex · fireworks/kimi-k3: 175k tokens per pass (15/22 passed)hard-review · task version 2 · codex · fireworks/kimi-k3175khard-review · task version 2 · codex · openai/gpt-6.1-sol @high: 230k tokens per pass (39/58 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol @high230khard-review · task version 2 · claude-code · openai/gpt-6.1-sol @high: 352k tokens per pass (6/9 passed)hard-review · task version 2 · claude-code · openai/gpt-6.1-sol @high352khard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @xhigh: 498k tokens per pass (6/9 passed)hard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @xhigh498khard-review · task version 2 · codex · anthropic/claude-opus-5-5: 136k tokens per pass (6/9 passed)hard-review · task version 2 · codex · anthropic/claude-opus-5-5136khard-review · task version 2 · codex · anthropic/claude-opus-5-5 @medium: 150k tokens per pass (6/9 passed)hard-review · task version 2 · codex · anthropic/claude-opus-5-5 @medium150khard-review · task version 2 · opencode · anthropic/claude-opus-5-5: 144k tokens per pass (6/9 passed)hard-review · task version 2 · opencode · anthropic/claude-opus-5-5144khard-review · task version 2 · codex · openai/gpt-6.1-sol @xhigh: 286k tokens per pass (21/33 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol @xhigh286khard-review · task version 2 · antigravity · google/gemini-3.8-flash @low: 881k tokens per pass (5/8 passed)hard-review · task version 2 · antigravity · google/gemini-3.8-flash @low881khard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @high: 1.31M tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @high1.31Mhard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @low: 766k tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @low766khard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @medium: 1.03M tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @medium1.03Mhard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @xhigh: 2.62M tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @xhigh2.62Mhard-review · task version 2 · claude-code · google/gemini-3.8-flash @high: 2.24M tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · google/gemini-3.8-flash @high2.24Mhard-review · task version 2 · claude-code · google/gemini-3.8-flash @low: 542k tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · google/gemini-3.8-flash @low542khard-review · task version 2 · claude-code · openai/gpt-6-luna @low: 412k tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · openai/gpt-6-luna @low412khard-review · task version 2 · claude-code · openai/gpt-6-luna @xhigh: 310k tokens per pass (5/8 passed)hard-review · task version 2 · claude-code · openai/gpt-6-luna @xhigh310khard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @low: 174k tokens per pass (5/8 passed)hard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @low174khard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @high: 160k tokens per pass (5/8 passed)hard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @high160khard-review · task version 2 · codex · google/gemini-3.8-flash @low: 784k tokens per pass (5/8 passed)hard-review · task version 2 · codex · google/gemini-3.8-flash @low784khard-review · task version 2 · codex · openai/gpt-5.6-terra @low: 322k tokens per pass (5/8 passed)hard-review · task version 2 · codex · openai/gpt-5.6-terra @low322khard-review · task version 2 · codex · openai/gpt-5.6-terra @max: 676k tokens per pass (5/8 passed)hard-review · task version 2 · codex · openai/gpt-5.6-terra @max676khard-review · task version 2 · codex · openai/gpt-6.1-sol @max: 350k tokens per pass (5/8 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol @max350khard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @high: 246k tokens per pass (5/8 passed)hard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @high246khard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @low: 175k tokens per pass (5/8 passed)hard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @low175khard-review · task version 2 · opencode · openai/gpt-5.6-terra @high: 334k tokens per pass (5/8 passed)hard-review · task version 2 · opencode · openai/gpt-5.6-terra @high334khard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @low: 110k tokens per pass (5/8 passed)hard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @low110khard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @medium: 131k tokens per pass (5/8 passed)hard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @medium131khard-review · task version 2 · pi · anthropic/claude-opus-5-5 @medium: 85k tokens per pass (5/8 passed)hard-review · task version 2 · pi · anthropic/claude-opus-5-5 @medium85khard-review · task version 2 · pi · openai/gpt-5.6-terra @high: 95k tokens per pass (5/8 passed)hard-review · task version 2 · pi · openai/gpt-5.6-terra @high95khard-review · task version 2 · pi · openai/gpt-5.6-terra @low: 72k tokens per pass (5/8 passed)hard-review · task version 2 · pi · openai/gpt-5.6-terra @low72khard-review · task version 2 · pi · openai/gpt-6-luna @max: 281k tokens per pass (5/8 passed)hard-review · task version 2 · pi · openai/gpt-6-luna @max281khard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5: 1.49M tokens per pass (6/10 passed)hard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-51.49Mhard-review · task version 2 · claude-code · google/gemini-3.8-flash: 2.42M tokens per pass (6/10 passed)hard-review · task version 2 · claude-code · google/gemini-3.8-flash2.42Mhard-review · task version 2 · codex · baseten/glm-5.3-flash: 271k tokens per pass (6/10 passed)hard-review · task version 2 · codex · baseten/glm-5.3-flash271khard-review · task version 2 · codex · moonshot/kimi-k3: 243k tokens per pass (6/10 passed)hard-review · task version 2 · codex · moonshot/kimi-k3243khard-review · task version 2 · pi · anthropic/claude-sonnet-5-5: 94k tokens per pass (6/10 passed)hard-review · task version 2 · pi · anthropic/claude-sonnet-5-594khard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @medium: 225k tokens per pass (5/9 passed)hard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @medium225khard-review · task version 2 · codex · openai/gpt-6.1-sol @ultra: 504k tokens per pass (5/9 passed)hard-review · task version 2 · codex · openai/gpt-6.1-sol @ultra504khard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @high: 270k tokens per pass (5/9 passed)hard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @high270khard-review · task version 2 · opencode · openai/gpt-6.1-sol @high: 406k tokens per pass (5/9 passed)hard-review · task version 2 · opencode · openai/gpt-6.1-sol @high406khard-review · task version 2 · opencode · openai/gpt-6.1-sol @max: 562k tokens per pass (5/9 passed)hard-review · task version 2 · opencode · openai/gpt-6.1-sol @max562khard-review · task version 2 · pi · anthropic/claude-opus-5-5: 105k tokens per pass (5/9 passed)hard-review · task version 2 · pi · anthropic/claude-opus-5-5105khard-review · task version 2 · pi · anthropic/claude-opus-5-5 @high: 110k tokens per pass (5/9 passed)hard-review · task version 2 · pi · anthropic/claude-opus-5-5 @high110khard-review · task version 2 · pi · openai/gpt-6.1-sol @high: 138k tokens per pass (5/9 passed)hard-review · task version 2 · pi · openai/gpt-6.1-sol @high138khard-review · task version 2 · pi · openai/gpt-6.1-sol @max: 161k tokens per pass (5/9 passed)hard-review · task version 2 · pi · openai/gpt-6.1-sol @max161khard-review · task version 2 · opencode · openai/gpt-6.1-sol: 399k tokens per pass (12/22 passed)hard-review · task version 2 · opencode · openai/gpt-6.1-sol399khard-review · task version 2 · claude-code · anthropic/claude-opus-5-5: 2.27M tokens per pass (17/34 passed)hard-review · task version 2 · claude-code · anthropic/claude-opus-5-52.27Mhard-review · task version 2 · codex · anthropic/claude-sonnet-5-5: 161k tokens per pass (5/10 passed)hard-review · task version 2 · codex · anthropic/claude-sonnet-5-5161khard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @max: 8.31M tokens per pass (4/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-haiku-5-5 @max8.31Mhard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @high: 1.91M tokens per pass (4/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @high1.91Mhard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @medium: 1.77M tokens per pass (4/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @medium1.77Mhard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @high: 1.87M tokens per pass (4/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @high1.87Mhard-review · task version 2 · claude-code · openai/gpt-6-luna @max: 401k tokens per pass (4/8 passed)hard-review · task version 2 · claude-code · openai/gpt-6-luna @max401khard-review · task version 2 · claude-code · openai/gpt-6.1-sol @max: 612k tokens per pass (4/8 passed)hard-review · task version 2 · claude-code · openai/gpt-6.1-sol @max612khard-review · task version 2 · codex · openai/gpt-6-luna @medium: 268k tokens per pass (4/8 passed)hard-review · task version 2 · codex · openai/gpt-6-luna @medium268khard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @xhigh: 795k tokens per pass (4/8 passed)hard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @xhigh795khard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @medium: 207k tokens per pass (4/8 passed)hard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @medium207khard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @high: 232k tokens per pass (4/8 passed)hard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @high232khard-review · task version 2 · opencode · google/gemini-3.8-flash @low: 422k tokens per pass (4/8 passed)hard-review · task version 2 · opencode · google/gemini-3.8-flash @low422khard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @medium: 94k tokens per pass (4/8 passed)hard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @medium94khard-review · task version 2 · pi · openai/gpt-6-luna @low: 77k tokens per pass (4/8 passed)hard-review · task version 2 · pi · openai/gpt-6-luna @low77khard-review · task version 2 · pi · openai/gpt-6-luna @xhigh: 198k tokens per pass (4/8 passed)hard-review · task version 2 · pi · openai/gpt-6-luna @xhigh198khard-review · task version 2 · antigravity · google/gemini-3.1-pro @high: 862k tokens per pass (4/9 passed)hard-review · task version 2 · antigravity · google/gemini-3.1-pro @high862khard-review · task version 2 · opencode · openai/gpt-6-luna @max: 759k tokens per pass (4/9 passed)hard-review · task version 2 · opencode · openai/gpt-6-luna @max759khard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @low: 2.29M tokens per pass (3/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @low2.29Mhard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @medium: 2.14M tokens per pass (3/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-sonnet-5-5 @medium2.14Mhard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @high: 604k tokens per pass (3/8 passed)hard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @high604khard-review · task version 2 · codex · openai/gpt-5.6-terra @ultra: 2.05M tokens per pass (3/8 passed)hard-review · task version 2 · codex · openai/gpt-5.6-terra @ultra2.05Mhard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @max: 3.49M tokens per pass (3/8 passed)hard-review · task version 2 · opencode · anthropic/claude-haiku-5-5 @max3.49Mhard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @max: 2.15M tokens per pass (3/8 passed)hard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @max2.15Mhard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @xhigh: 900k tokens per pass (3/8 passed)hard-review · task version 2 · pi · anthropic/claude-haiku-5-5 @xhigh900khard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @max: 2.28M tokens per pass (3/9 passed)hard-review · task version 2 · codex · anthropic/claude-haiku-5-5 @max2.28Mhard-review · task version 2 · codex · anthropic/claude-opus-5-5 @low: 244k tokens per pass (3/9 passed)hard-review · task version 2 · codex · anthropic/claude-opus-5-5 @low244khard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @low: 281k tokens per pass (3/9 passed)hard-review · task version 2 · opencode · anthropic/claude-opus-5-5 @low281khard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @low: 3.25M tokens per pass (2/8 passed)hard-review · task version 2 · claude-code · anthropic/claude-opus-5-5 @low3.25Mhard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @low: 353k tokens per pass (2/8 passed)hard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @low353khard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @medium: 461k tokens per pass (2/8 passed)hard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @medium461khard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @low: 172k tokens per pass (2/8 passed)hard-review · task version 2 · pi · anthropic/claude-sonnet-5-5 @low172khard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @medium: 450k tokens per pass (2/9 passed)hard-review · task version 2 · codex · anthropic/claude-sonnet-5-5 @medium450khard-review · task version 2 · pi · anthropic/claude-opus-5-5 @low: 206k tokens per pass (2/9 passed)hard-review · task version 2 · pi · anthropic/claude-opus-5-5 @low206khard-review · task version 2 · codex · openai/gpt-6-luna @low: 1.00M tokens per pass (1/8 passed)hard-review · task version 2 · codex · openai/gpt-6-luna @low1.00Mhard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @low: 921k tokens per pass (1/8 passed)hard-review · task version 2 · opencode · anthropic/claude-sonnet-5-5 @low921khard-review · task version 3 · codex · openai/gpt-6.1-sol @medium: 134k tokens per pass (11/11 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol @medium134khard-review · task version 3 · codex · openai/gpt-6.1-sol @high: 161k tokens per pass (10/10 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol @high161khard-review · task version 3 · claude-code · anthropic/claude-opus-5-5: 1.05M tokens per pass (5/5 passed)hard-review · task version 3 · claude-code · anthropic/claude-opus-5-51.05Mhard-review · task version 3 · codex · openai/gpt-6.1-sol: 132k tokens per pass (5/5 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol132khard-review · task version 3 · codex · openai/gpt-6.1-sol @xhigh: 170k tokens per pass (5/5 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol @xhigh170khard-review · task version 3 · antigravity · google/gemini-3.8-flash @high: 972k tokens per pass (3/3 passed)hard-review · task version 3 · antigravity · google/gemini-3.8-flash @high972khard-review · task version 3 · claude-code · together/deepseek-v4.1-flash: 337k tokens per pass (3/3 passed)hard-review · task version 3 · claude-code · together/deepseek-v4.1-flash337khard-review · task version 3 · codex · fireworks/kimi-k3: 85k tokens per pass (3/3 passed)hard-review · task version 3 · codex · fireworks/kimi-k385khard-review · task version 3 · codex · zai/glm-5.2: 552k tokens per pass (3/3 passed)hard-review · task version 3 · codex · zai/glm-5.2552khard-review · task version 3 · opencode · baseten/glm-5.3-flash: 165k tokens per pass (3/3 passed)hard-review · task version 3 · opencode · baseten/glm-5.3-flash165khard-review · task version 3 · opencode · openai/gpt-6.1-sol: 190k tokens per pass (3/3 passed)hard-review · task version 3 · opencode · openai/gpt-6.1-sol190khard-review · task version 3 · pi · baseten/glm-5.3-flash: 143k tokens per pass (3/3 passed)hard-review · task version 3 · pi · baseten/glm-5.3-flash143khard-review · task version 3 · antigravity · google/gemini-3.1-pro @high: 360k tokens per pass (2/2 passed)hard-review · task version 3 · antigravity · google/gemini-3.1-pro @high360khard-review · task version 3 · claude-code · zai/glm-5.2: 538k tokens per pass (2/2 passed)hard-review · task version 3 · claude-code · zai/glm-5.2538khard-review · task version 3 · codex · together/deepseek-v4.1-flash: 117k tokens per pass (2/2 passed)hard-review · task version 3 · codex · together/deepseek-v4.1-flash117khard-review · task version 3 · opencode · together/deepseek-v4.1-flash: 156k tokens per pass (2/2 passed)hard-review · task version 3 · opencode · together/deepseek-v4.1-flash156khard-review · task version 3 · antigravity · google/gemini-3.8-flash @medium: 2.41M tokens per pass (1/1 passed)hard-review · task version 3 · antigravity · google/gemini-3.8-flash @medium2.41Mhard-review · task version 3 · antigravity · google/gemini-4-argon @low: 85k tokens per pass (1/1 passed)hard-review · task version 3 · antigravity · google/gemini-4-argon @low85khard-review · task version 3 · claude-code · anthropic/claude-haiku-5-5 @low: 430k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-haiku-5-5 @low430khard-review · task version 3 · claude-code · anthropic/claude-haiku-5-5 @medium: 366k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-haiku-5-5 @medium366khard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @high: 771k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @high771khard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @max: 4.18M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @max4.18Mhard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @medium: 1.26M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @medium1.26Mhard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5: 1.04M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-51.04Mhard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @high: 1.20M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @high1.20Mhard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @low: 897k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @low897khard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @max: 6.26M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @max6.26Mhard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @medium: 737k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · anthropic/claude-sonnet-5-5 @medium737khard-review · task version 3 · claude-code · fireworks/kimi-k3: 266k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · fireworks/kimi-k3266khard-review · task version 3 · claude-code · google/gemini-3.8-flash @high: 1.43M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · google/gemini-3.8-flash @high1.43Mhard-review · task version 3 · claude-code · google/gemini-3.8-flash @low: 4.81M tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · google/gemini-3.8-flash @low4.81Mhard-review · task version 3 · claude-code · google/gemini-3.8-flash @medium: 835k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · google/gemini-3.8-flash @medium835khard-review · task version 3 · claude-code · moonshot/kimi-k3: 274k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · moonshot/kimi-k3274khard-review · task version 3 · claude-code · openai/gpt-5.6-terra @high: 128k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-5.6-terra @high128khard-review · task version 3 · claude-code · openai/gpt-5.6-terra @low: 146k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-5.6-terra @low146khard-review · task version 3 · claude-code · openai/gpt-6-luna @high: 153k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-6-luna @high153khard-review · task version 3 · claude-code · openai/gpt-6-luna @max: 133k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-6-luna @max133khard-review · task version 3 · claude-code · openai/gpt-6.1-sol: 259k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-6.1-sol259khard-review · task version 3 · claude-code · openai/gpt-6.1-sol @high: 176k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-6.1-sol @high176khard-review · task version 3 · claude-code · openai/gpt-6.1-sol @low: 244k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-6.1-sol @low244khard-review · task version 3 · claude-code · openai/gpt-6.1-sol @max: 227k tokens per pass (1/1 passed)hard-review · task version 3 · claude-code · openai/gpt-6.1-sol @max227khard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @high: 139k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @high139khard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @low: 133k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @low133khard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @medium: 94k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @medium94khard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @xhigh: 285k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-haiku-5-5 @xhigh285khard-review · task version 3 · codex · anthropic/claude-opus-5-5: 91k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-opus-5-591khard-review · task version 3 · codex · anthropic/claude-opus-5-5 @high: 116k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-opus-5-5 @high116khard-review · task version 3 · codex · anthropic/claude-opus-5-5 @low: 108k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-opus-5-5 @low108khard-review · task version 3 · codex · anthropic/claude-opus-5-5 @medium: 117k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-opus-5-5 @medium117khard-review · task version 3 · codex · anthropic/claude-opus-5-5 @xhigh: 246k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-opus-5-5 @xhigh246khard-review · task version 3 · codex · anthropic/claude-sonnet-5-5 @high: 86k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-sonnet-5-5 @high86khard-review · task version 3 · codex · anthropic/claude-sonnet-5-5 @medium: 62k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-sonnet-5-5 @medium62khard-review · task version 3 · codex · anthropic/claude-sonnet-5-5 @xhigh: 114k tokens per pass (1/1 passed)hard-review · task version 3 · codex · anthropic/claude-sonnet-5-5 @xhigh114khard-review · task version 3 · codex · fireworks/qwen3.8-max: 204k tokens per pass (1/1 passed)hard-review · task version 3 · codex · fireworks/qwen3.8-max204khard-review · task version 3 · codex · google/gemini-3.8-flash: 1.01M tokens per pass (1/1 passed)hard-review · task version 3 · codex · google/gemini-3.8-flash1.01Mhard-review · task version 3 · codex · google/gemini-3.8-flash @medium: 591k tokens per pass (1/1 passed)hard-review · task version 3 · codex · google/gemini-3.8-flash @medium591khard-review · task version 3 · codex · openai/gpt-5.6-terra @low: 144k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-5.6-terra @low144khard-review · task version 3 · codex · openai/gpt-5.6-terra @xhigh: 255k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-5.6-terra @xhigh255khard-review · task version 3 · codex · openai/gpt-6-luna @high: 130k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6-luna @high130khard-review · task version 3 · codex · openai/gpt-6-luna @max: 159k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6-luna @max159khard-review · task version 3 · codex · openai/gpt-6-luna @medium: 124k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6-luna @medium124khard-review · task version 3 · codex · openai/gpt-6-luna @xhigh: 202k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6-luna @xhigh202khard-review · task version 3 · codex · openai/gpt-6.1-sol @low: 108k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol @low108khard-review · task version 3 · codex · openai/gpt-6.1-sol @max: 255k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol @max255khard-review · task version 3 · codex · openai/gpt-6.1-sol @ultra: 294k tokens per pass (1/1 passed)hard-review · task version 3 · codex · openai/gpt-6.1-sol @ultra294khard-review · task version 3 · opencode · anthropic/claude-haiku-5-5 @high: 111k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-haiku-5-5 @high111khard-review · task version 3 · opencode · anthropic/claude-haiku-5-5 @low: 94k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-haiku-5-5 @low94khard-review · task version 3 · opencode · anthropic/claude-haiku-5-5 @xhigh: 243k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-haiku-5-5 @xhigh243khard-review · task version 3 · opencode · anthropic/claude-opus-5-5: 79k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-opus-5-579khard-review · task version 3 · opencode · anthropic/claude-opus-5-5 @high: 142k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-opus-5-5 @high142khard-review · task version 3 · opencode · anthropic/claude-opus-5-5 @medium: 98k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-opus-5-5 @medium98khard-review · task version 3 · opencode · anthropic/claude-opus-5-5 @xhigh: 203k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-opus-5-5 @xhigh203khard-review · task version 3 · opencode · anthropic/claude-sonnet-5-5: 90k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-sonnet-5-590khard-review · task version 3 · opencode · anthropic/claude-sonnet-5-5 @high: 99k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-sonnet-5-5 @high99khard-review · task version 3 · opencode · anthropic/claude-sonnet-5-5 @xhigh: 140k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · anthropic/claude-sonnet-5-5 @xhigh140khard-review · task version 3 · opencode · fireworks/kimi-k3: 107k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · fireworks/kimi-k3107khard-review · task version 3 · opencode · fireworks/qwen3.8-max: 139k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · fireworks/qwen3.8-max139khard-review · task version 3 · opencode · google/gemini-3.8-flash: 809k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · google/gemini-3.8-flash809khard-review · task version 3 · opencode · google/gemini-3.8-flash @high: 691k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · google/gemini-3.8-flash @high691khard-review · task version 3 · opencode · google/gemini-3.8-flash @medium: 587k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · google/gemini-3.8-flash @medium587khard-review · task version 3 · opencode · moonshot/kimi-k3: 140k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · moonshot/kimi-k3140khard-review · task version 3 · opencode · openai/gpt-5.6-terra @high: 147k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-5.6-terra @high147khard-review · task version 3 · opencode · openai/gpt-5.6-terra @low: 54k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-5.6-terra @low54khard-review · task version 3 · opencode · openai/gpt-6-luna @high: 305k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-6-luna @high305khard-review · task version 3 · opencode · openai/gpt-6-luna @low: 146k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-6-luna @low146khard-review · task version 3 · opencode · openai/gpt-6-luna @xhigh: 177k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-6-luna @xhigh177khard-review · task version 3 · opencode · openai/gpt-6.1-sol @high: 215k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-6.1-sol @high215khard-review · task version 3 · opencode · openai/gpt-6.1-sol @low: 83k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-6.1-sol @low83khard-review · task version 3 · opencode · openai/gpt-6.1-sol @max: 316k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · openai/gpt-6.1-sol @max316khard-review · task version 3 · opencode · zai/glm-5.2: 199k tokens per pass (1/1 passed)hard-review · task version 3 · opencode · zai/glm-5.2199khard-review · task version 3 · pi · anthropic/claude-haiku-5-5 @high: 62k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-haiku-5-5 @high62khard-review · task version 3 · pi · anthropic/claude-haiku-5-5 @low: 56k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-haiku-5-5 @low56khard-review · task version 3 · pi · anthropic/claude-opus-5-5: 58k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-opus-5-558khard-review · task version 3 · pi · anthropic/claude-opus-5-5 @high: 53k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-opus-5-5 @high53khard-review · task version 3 · pi · anthropic/claude-opus-5-5 @low: 35k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-opus-5-5 @low35khard-review · task version 3 · pi · anthropic/claude-opus-5-5 @medium: 60k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-opus-5-5 @medium60khard-review · task version 3 · pi · anthropic/claude-opus-5-5 @xhigh: 115k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-opus-5-5 @xhigh115khard-review · task version 3 · pi · anthropic/claude-sonnet-5-5 @high: 43k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-sonnet-5-5 @high43khard-review · task version 3 · pi · anthropic/claude-sonnet-5-5 @medium: 46k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-sonnet-5-5 @medium46khard-review · task version 3 · pi · anthropic/claude-sonnet-5-5 @xhigh: 55k tokens per pass (1/1 passed)hard-review · task version 3 · pi · anthropic/claude-sonnet-5-5 @xhigh55khard-review · task version 3 · pi · fireworks/kimi-k3: 121k tokens per pass (1/1 passed)hard-review · task version 3 · pi · fireworks/kimi-k3121khard-review · task version 3 · pi · google/gemini-3.8-flash @high: 782k tokens per pass (1/1 passed)hard-review · task version 3 · pi · google/gemini-3.8-flash @high782khard-review · task version 3 · pi · google/gemini-3.8-flash @medium: 879k tokens per pass (1/1 passed)hard-review · task version 3 · pi · google/gemini-3.8-flash @medium879khard-review · task version 3 · pi · moonshot/kimi-k3: 109k tokens per pass (1/1 passed)hard-review · task version 3 · pi · moonshot/kimi-k3109khard-review · task version 3 · pi · openai/gpt-5.6-terra @low: 33k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-5.6-terra @low33khard-review · task version 3 · pi · openai/gpt-6-luna @high: 43k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6-luna @high43khard-review · task version 3 · pi · openai/gpt-6-luna @max: 101k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6-luna @max101khard-review · task version 3 · pi · openai/gpt-6-luna @xhigh: 60k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6-luna @xhigh60khard-review · task version 3 · pi · openai/gpt-6.1-sol: 75k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6.1-sol75khard-review · task version 3 · pi · openai/gpt-6.1-sol @high: 66k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6.1-sol @high66khard-review · task version 3 · pi · openai/gpt-6.1-sol @low: 43k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6.1-sol @low43khard-review · task version 3 · pi · openai/gpt-6.1-sol @max: 85k tokens per pass (1/1 passed)hard-review · task version 3 · pi · openai/gpt-6.1-sol @max85khard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @xhigh: 1.78M tokens per pass (4/5 passed)hard-review · task version 3 · claude-code · anthropic/claude-opus-5-5 @xhigh1.78Mhard-review · task version 3 · pi · together/deepseek-v4.1-flash: 198k tokens per pass (4/5 passed)hard-review · task version 3 · pi · together/deepseek-v4.1-flash198khard-review · task version 3 · claude-code · fireworks/qwen3.8-max: 348k tokens per pass (3/4 passed)hard-review · task version 3 · claude-code · fireworks/qwen3.8-max348khard-review · task version 3 · pi · google/gemini-3.8-flash: 844k tokens per pass (3/4 passed)hard-review · task version 3 · pi · google/gemini-3.8-flash844khard-review · task version 3 · pi · fireworks/qwen3.8-max: 177k tokens per pass (2/3 passed)hard-review · task version 3 · pi · fireworks/qwen3.8-max177krca · codex · openai/gpt-6.1-sol @medium: 66k tokens per pass (14/14 passed)rca · codex · openai/gpt-6.1-sol @medium66krca · codex · openai/gpt-6.1-sol @high: 67k tokens per pass (12/12 passed)rca · codex · openai/gpt-6.1-sol @high67krca · claude-code · fireworks/qwen3.8-max: 146k tokens per pass (7/7 passed)rca · claude-code · fireworks/qwen3.8-max146krca · pi · google/gemini-3.8-flash: 85k tokens per pass (7/7 passed)rca · pi · google/gemini-3.8-flash85krca · pi · together/deepseek-v4.1-flash: 17k tokens per pass (7/7 passed)rca · pi · together/deepseek-v4.1-flash17krca · opencode · together/deepseek-v4.1-flash: 38k tokens per pass (6/6 passed)rca · opencode · together/deepseek-v4.1-flash38krca · claude-code · anthropic/claude-opus-5-5 @xhigh: 637k tokens per pass (5/5 passed)rca · claude-code · anthropic/claude-opus-5-5 @xhigh637krca · claude-code · together/deepseek-v4.1-flash: 175k tokens per pass (5/5 passed)rca · claude-code · together/deepseek-v4.1-flash175krca · codex · fireworks/kimi-k3: 33k tokens per pass (5/5 passed)rca · codex · fireworks/kimi-k333krca · codex · zai/glm-5.2: 54k tokens per pass (5/5 passed)rca · codex · zai/glm-5.254krca · codex · openai/gpt-6.1-sol @xhigh: 71k tokens per pass (5/5 passed)rca · codex · openai/gpt-6.1-sol @xhigh71krca · opencode · baseten/glm-5.3-flash: 32k tokens per pass (5/5 passed)rca · opencode · baseten/glm-5.3-flash32krca · opencode · openai/gpt-6.1-sol: 54k tokens per pass (5/5 passed)rca · opencode · openai/gpt-6.1-sol54krca · pi · baseten/glm-5.3-flash: 13k tokens per pass (5/5 passed)rca · pi · baseten/glm-5.3-flash13krca · antigravity · google/gemini-3.1-pro @high: 82k tokens per pass (3/3 passed)rca · antigravity · google/gemini-3.1-pro @high82krca · antigravity · google/gemini-3.8-flash @high: 410k tokens per pass (3/3 passed)rca · antigravity · google/gemini-3.8-flash @high410krca · antigravity · google/gemini-4-argon @high: 102k tokens per pass (3/3 passed)rca · antigravity · google/gemini-4-argon @high102krca · claude-code · anthropic/claude-sonnet-5-5: 506k tokens per pass (3/3 passed)rca · claude-code · anthropic/claude-sonnet-5-5506krca · claude-code · baseten/glm-5.3-flash: 182k tokens per pass (3/3 passed)rca · claude-code · baseten/glm-5.3-flash182krca · claude-code · google/gemini-3.8-flash: 402k tokens per pass (3/3 passed)rca · claude-code · google/gemini-3.8-flash402krca · claude-code · google/gemini-3.8-flash @high: 1.09M tokens per pass (3/3 passed)rca · claude-code · google/gemini-3.8-flash @high1.09Mrca · claude-code · google/gemini-3.8-flash @low: 67k tokens per pass (3/3 passed)rca · claude-code · google/gemini-3.8-flash @low67krca · claude-code · google/gemini-3.8-flash @medium: 398k tokens per pass (3/3 passed)rca · claude-code · google/gemini-3.8-flash @medium398krca · claude-code · moonshot/kimi-k3: 123k tokens per pass (3/3 passed)rca · claude-code · moonshot/kimi-k3123krca · claude-code · openai/gpt-6.1-sol: 123k tokens per pass (3/3 passed)rca · claude-code · openai/gpt-6.1-sol123krca · codex · anthropic/claude-opus-5-5: 40k tokens per pass (3/3 passed)rca · codex · anthropic/claude-opus-5-540krca · codex · anthropic/claude-sonnet-5-5: 38k tokens per pass (3/3 passed)rca · codex · anthropic/claude-sonnet-5-538krca · codex · baseten/glm-5.3-flash: 25k tokens per pass (3/3 passed)rca · codex · baseten/glm-5.3-flash25krca · codex · google/gemini-3.8-flash: 188k tokens per pass (3/3 passed)rca · codex · google/gemini-3.8-flash188krca · codex · google/gemini-3.8-flash @low: 57k tokens per pass (3/3 passed)rca · codex · google/gemini-3.8-flash @low57krca · codex · moonshot/kimi-k3: 33k tokens per pass (3/3 passed)rca · codex · moonshot/kimi-k333krca · codex · together/deepseek-v4.1-flash: 30k tokens per pass (3/3 passed)rca · codex · together/deepseek-v4.1-flash30krca · opencode · anthropic/claude-opus-5-5: 47k tokens per pass (3/3 passed)rca · opencode · anthropic/claude-opus-5-547krca · opencode · anthropic/claude-sonnet-5-5: 44k tokens per pass (3/3 passed)rca · opencode · anthropic/claude-sonnet-5-544krca · opencode · fireworks/kimi-k3: 56k tokens per pass (3/3 passed)rca · opencode · fireworks/kimi-k356krca · opencode · google/gemini-3.8-flash: 167k tokens per pass (3/3 passed)rca · opencode · google/gemini-3.8-flash167krca · opencode · google/gemini-3.8-flash @high: 451k tokens per pass (3/3 passed)rca · opencode · google/gemini-3.8-flash @high451krca · opencode · google/gemini-3.8-flash @low: 76k tokens per pass (3/3 passed)rca · opencode · google/gemini-3.8-flash @low76krca · opencode · moonshot/kimi-k3: 61k tokens per pass (3/3 passed)rca · opencode · moonshot/kimi-k361krca · opencode · zai/glm-5.2: 49k tokens per pass (3/3 passed)rca · opencode · zai/glm-5.249krca · pi · anthropic/claude-opus-5-5: 18k tokens per pass (3/3 passed)rca · pi · anthropic/claude-opus-5-518krca · pi · anthropic/claude-sonnet-5-5: 17k tokens per pass (3/3 passed)rca · pi · anthropic/claude-sonnet-5-517krca · pi · fireworks/kimi-k3: 20k tokens per pass (3/3 passed)rca · pi · fireworks/kimi-k320krca · pi · google/gemini-3.8-flash @high: 216k tokens per pass (3/3 passed)rca · pi · google/gemini-3.8-flash @high216krca · pi · moonshot/kimi-k3: 23k tokens per pass (3/3 passed)rca · pi · moonshot/kimi-k323krca · pi · openai/gpt-6.1-sol: 16k tokens per pass (3/3 passed)rca · pi · openai/gpt-6.1-sol16krca · pi · zai/glm-5.2: 27k tokens per pass (3/3 passed)rca · pi · zai/glm-5.227krca · codex · openai/gpt-5.6-luna @medium: 55k tokens per pass (1/1 passed)rca · codex · openai/gpt-5.6-luna @medium55krca · codex · openai/gpt-5.6-luna @xhigh: 55k tokens per pass (1/1 passed)rca · codex · openai/gpt-5.6-luna @xhigh55krca · codex · openai/gpt-5.6-sol @high: 61k tokens per pass (1/1 passed)rca · codex · openai/gpt-5.6-sol @high61krca · pi · google/gemini-3.8-flash @low: 12k tokens per pass (1/1 passed)rca · pi · google/gemini-3.8-flash @low12krca · opencode · fireworks/qwen3.8-max: 55k tokens per pass (5/6 passed)rca · opencode · fireworks/qwen3.8-max55krca · codex · openai/gpt-6.1-sol: 81k tokens per pass (9/11 passed)rca · codex · openai/gpt-6.1-sol81krca · claude-code · zai/glm-5.2: 223k tokens per pass (7/9 passed)rca · claude-code · zai/glm-5.2223krca · pi · fireworks/qwen3.8-max: 45k tokens per pass (5/7 passed)rca · pi · fireworks/qwen3.8-max45krca · claude-code · anthropic/claude-haiku-4-5: 724k tokens per pass (2/3 passed)rca · claude-code · anthropic/claude-haiku-4-5724krca · claude-code · anthropic/claude-sonnet-5: 975k tokens per pass (2/3 passed)rca · claude-code · anthropic/claude-sonnet-5975krca · claude-code · fireworks/kimi-k3: 197k tokens per pass (2/3 passed)rca · claude-code · fireworks/kimi-k3197krca · claude-code · anthropic/claude-opus-5-5: 1.06M tokens per pass (7/11 passed)rca · claude-code · anthropic/claude-opus-5-51.06Mrca · codex · fireworks/qwen3.8-max: 130k tokens per pass (3/6 passed)rca · codex · fireworks/qwen3.8-max130kreview · task version 1 · antigravity · google/gemini-3.8-flash @high: 350k tokens per pass (9/9 passed)review · task version 1 · antigravity · google/gemini-3.8-flash @high350kreview · task version 1 · opencode · together/deepseek-v4.1-flash: 47k tokens per pass (8/8 passed)review · task version 1 · opencode · together/deepseek-v4.1-flash47kreview · task version 1 · claude-code · together/deepseek-v4.1-flash: 217k tokens per pass (6/6 passed)review · task version 1 · claude-code · together/deepseek-v4.1-flash217kreview · task version 1 · antigravity · google/gemini-4-argon @high: 104k tokens per pass (3/3 passed)review · task version 1 · antigravity · google/gemini-4-argon @high104kreview · task version 1 · claude-code · baseten/glm-5.3-flash: 234k tokens per pass (3/3 passed)review · task version 1 · claude-code · baseten/glm-5.3-flash234kreview · task version 1 · claude-code · fireworks/kimi-k3: 136k tokens per pass (3/3 passed)review · task version 1 · claude-code · fireworks/kimi-k3136kreview · task version 1 · claude-code · google/gemini-3.8-flash: 417k tokens per pass (3/3 passed)review · task version 1 · claude-code · google/gemini-3.8-flash417kreview · task version 1 · claude-code · moonshot/kimi-k3: 176k tokens per pass (3/3 passed)review · task version 1 · claude-code · moonshot/kimi-k3176kreview · task version 1 · claude-code · openai/gpt-6.1-sol: 145k tokens per pass (3/3 passed)review · task version 1 · claude-code · openai/gpt-6.1-sol145kreview · task version 1 · codex · anthropic/claude-opus-5-5: 51k tokens per pass (3/3 passed)review · task version 1 · codex · anthropic/claude-opus-5-551kreview · task version 1 · codex · anthropic/claude-sonnet-5-5: 44k tokens per pass (3/3 passed)review · task version 1 · codex · anthropic/claude-sonnet-5-544kreview · task version 1 · codex · fireworks/kimi-k3: 31k tokens per pass (3/3 passed)review · task version 1 · codex · fireworks/kimi-k331kreview · task version 1 · codex · fireworks/qwen3.8-max: 56k tokens per pass (3/3 passed)review · task version 1 · codex · fireworks/qwen3.8-max56kreview · task version 1 · codex · google/gemini-3.8-flash: 346k tokens per pass (3/3 passed)review · task version 1 · codex · google/gemini-3.8-flash346kreview · task version 1 · codex · moonshot/kimi-k3: 36k tokens per pass (3/3 passed)review · task version 1 · codex · moonshot/kimi-k336kreview · task version 1 · codex · openai/gpt-6.1-sol @high: 76k tokens per pass (3/3 passed)review · task version 1 · codex · openai/gpt-6.1-sol @high76kreview · task version 1 · opencode · anthropic/claude-opus-5-5: 56k tokens per pass (3/3 passed)review · task version 1 · opencode · anthropic/claude-opus-5-556kreview · task version 1 · opencode · anthropic/claude-sonnet-5-5: 57k tokens per pass (3/3 passed)review · task version 1 · opencode · anthropic/claude-sonnet-5-557kreview · task version 1 · opencode · baseten/glm-5.3-flash: 48k tokens per pass (3/3 passed)review · task version 1 · opencode · baseten/glm-5.3-flash48kreview · task version 1 · opencode · fireworks/kimi-k3: 59k tokens per pass (3/3 passed)review · task version 1 · opencode · fireworks/kimi-k359kreview · task version 1 · opencode · fireworks/qwen3.8-max: 69k tokens per pass (3/3 passed)review · task version 1 · opencode · fireworks/qwen3.8-max69kreview · task version 1 · opencode · google/gemini-3.8-flash: 117k tokens per pass (3/3 passed)review · task version 1 · opencode · google/gemini-3.8-flash117kreview · task version 1 · opencode · moonshot/kimi-k3: 59k tokens per pass (3/3 passed)review · task version 1 · opencode · moonshot/kimi-k359kreview · task version 1 · opencode · openai/gpt-6.1-sol: 62k tokens per pass (3/3 passed)review · task version 1 · opencode · openai/gpt-6.1-sol62kreview · task version 1 · pi · anthropic/claude-opus-5-5: 22k tokens per pass (3/3 passed)review · task version 1 · pi · anthropic/claude-opus-5-522kreview · task version 1 · pi · anthropic/claude-sonnet-5-5: 18k tokens per pass (3/3 passed)review · task version 1 · pi · anthropic/claude-sonnet-5-518kreview · task version 1 · pi · baseten/glm-5.3-flash: 17k tokens per pass (3/3 passed)review · task version 1 · pi · baseten/glm-5.3-flash17kreview · task version 1 · pi · fireworks/kimi-k3: 18k tokens per pass (3/3 passed)review · task version 1 · pi · fireworks/kimi-k318kreview · task version 1 · pi · fireworks/qwen3.8-max: 41k tokens per pass (3/3 passed)review · task version 1 · pi · fireworks/qwen3.8-max41kreview · task version 1 · pi · google/gemini-3.8-flash: 111k tokens per pass (3/3 passed)review · task version 1 · pi · google/gemini-3.8-flash111kreview · task version 1 · pi · google/gemini-3.8-flash @high: 188k tokens per pass (3/3 passed)review · task version 1 · pi · google/gemini-3.8-flash @high188kreview · task version 1 · pi · moonshot/kimi-k3: 23k tokens per pass (3/3 passed)review · task version 1 · pi · moonshot/kimi-k323kreview · task version 1 · pi · openai/gpt-6.1-sol: 19k tokens per pass (3/3 passed)review · task version 1 · pi · openai/gpt-6.1-sol19kreview · task version 1 · codex · zai/glm-5.2: 315k tokens per pass (5/6 passed)review · task version 1 · codex · zai/glm-5.2315kreview · task version 1 · pi · zai/glm-5.2: 236k tokens per pass (4/6 passed)review · task version 1 · pi · zai/glm-5.2236kreview · task version 1 · codex · baseten/glm-5.3-flash: 60k tokens per pass (2/3 passed)review · task version 1 · codex · baseten/glm-5.3-flash60kreview · task version 1 · opencode · zai/glm-5.2: 243k tokens per pass (3/6 passed)review · task version 1 · opencode · zai/glm-5.2243kreview · task version 2 · codex · openai/gpt-6.1-sol @medium: 103k tokens per pass (23/23 passed)review · task version 2 · codex · openai/gpt-6.1-sol @medium103kreview · task version 2 · antigravity · google/gemini-3.8-flash @high: 589k tokens per pass (18/18 passed)review · task version 2 · antigravity · google/gemini-3.8-flash @high589kreview · task version 2 · codex · openai/gpt-6.1-sol @high: 101k tokens per pass (14/14 passed)review · task version 2 · codex · openai/gpt-6.1-sol @high101kreview · task version 2 · claude-code · anthropic/claude-opus-5-5: 757k tokens per pass (12/12 passed)review · task version 2 · claude-code · anthropic/claude-opus-5-5757kreview · task version 2 · codex · openai/gpt-6.1-sol: 94k tokens per pass (12/12 passed)review · task version 2 · codex · openai/gpt-6.1-sol94kreview · task version 2 · codex · openai/gpt-6.1-sol @xhigh: 114k tokens per pass (10/10 passed)review · task version 2 · codex · openai/gpt-6.1-sol @xhigh114kreview · task version 2 · opencode · openai/gpt-6.1-sol: 109k tokens per pass (10/10 passed)review · task version 2 · opencode · openai/gpt-6.1-sol109kreview · task version 2 · pi · together/deepseek-v4.1-flash: 89k tokens per pass (10/10 passed)review · task version 2 · pi · together/deepseek-v4.1-flash89kreview · task version 2 · opencode · together/deepseek-v4.1-flash: 60k tokens per pass (9/9 passed)review · task version 2 · opencode · together/deepseek-v4.1-flash60kreview · task version 2 · antigravity · google/gemini-3.1-pro @high: 278k tokens per pass (6/6 passed)review · task version 2 · antigravity · google/gemini-3.1-pro @high278kreview · task version 2 · antigravity · google/gemini-4-argon @high: 176k tokens per pass (6/6 passed)review · task version 2 · antigravity · google/gemini-4-argon @high176kreview · task version 2 · claude-code · google/gemini-3.8-flash: 610k tokens per pass (6/6 passed)review · task version 2 · claude-code · google/gemini-3.8-flash610kreview · task version 2 · claude-code · openai/gpt-6.1-sol: 169k tokens per pass (6/6 passed)review · task version 2 · claude-code · openai/gpt-6.1-sol169kreview · task version 2 · codex · google/gemini-3.8-flash: 423k tokens per pass (6/6 passed)review · task version 2 · codex · google/gemini-3.8-flash423kreview · task version 2 · opencode · anthropic/claude-sonnet-5-5: 67k tokens per pass (6/6 passed)review · task version 2 · opencode · anthropic/claude-sonnet-5-567kreview · task version 2 · opencode · google/gemini-3.8-flash: 283k tokens per pass (6/6 passed)review · task version 2 · opencode · google/gemini-3.8-flash283kreview · task version 2 · pi · anthropic/claude-sonnet-5-5: 22k tokens per pass (6/6 passed)review · task version 2 · pi · anthropic/claude-sonnet-5-522kreview · task version 2 · pi · fireworks/kimi-k3: 42k tokens per pass (6/6 passed)review · task version 2 · pi · fireworks/kimi-k342kreview · task version 2 · pi · google/gemini-3.8-flash: 273k tokens per pass (6/6 passed)review · task version 2 · pi · google/gemini-3.8-flash273kreview · task version 2 · pi · google/gemini-3.8-flash @high: 440k tokens per pass (6/6 passed)review · task version 2 · pi · google/gemini-3.8-flash @high440kreview · task version 2 · pi · moonshot/kimi-k3: 41k tokens per pass (6/6 passed)review · task version 2 · pi · moonshot/kimi-k341kreview · task version 2 · pi · openai/gpt-6.1-sol: 29k tokens per pass (6/6 passed)review · task version 2 · pi · openai/gpt-6.1-sol29kreview · task version 2 · claude-code · fireworks/qwen3.8-27b: 192k tokens per pass (3/3 passed)review · task version 2 · claude-code · fireworks/qwen3.8-27b192kreview · task version 2 · claude-code · google/gemini-3.8-flash @high: 523k tokens per pass (3/3 passed)review · task version 2 · claude-code · google/gemini-3.8-flash @high523kreview · task version 2 · claude-code · google/gemini-3.8-flash @low: 118k tokens per pass (3/3 passed)review · task version 2 · claude-code · google/gemini-3.8-flash @low118kreview · task version 2 · claude-code · google/gemini-3.8-flash @medium: 209k tokens per pass (3/3 passed)review · task version 2 · claude-code · google/gemini-3.8-flash @medium209kreview · task version 2 · codex · google/gemini-3.8-flash @low: 82k tokens per pass (3/3 passed)review · task version 2 · codex · google/gemini-3.8-flash @low82kreview · task version 2 · opencode · google/gemini-3.8-flash @high: 336k tokens per pass (3/3 passed)review · task version 2 · opencode · google/gemini-3.8-flash @high336kreview · task version 2 · opencode · google/gemini-3.8-flash @low: 87k tokens per pass (3/3 passed)review · task version 2 · opencode · google/gemini-3.8-flash @low87kreview · task version 2 · claude-code · together/deepseek-v4.1-flash: 268k tokens per pass (7/8 passed)review · task version 2 · claude-code · together/deepseek-v4.1-flash268kreview · task version 2 · claude-code · anthropic/claude-sonnet-5-5: 930k tokens per pass (5/6 passed)review · task version 2 · claude-code · anthropic/claude-sonnet-5-5930kreview · task version 2 · claude-code · baseten/glm-5.3-flash: 435k tokens per pass (5/6 passed)review · task version 2 · claude-code · baseten/glm-5.3-flash435kreview · task version 2 · claude-code · fireworks/kimi-k3: 274k tokens per pass (5/6 passed)review · task version 2 · claude-code · fireworks/kimi-k3274kreview · task version 2 · claude-code · moonshot/kimi-k3: 260k tokens per pass (5/6 passed)review · task version 2 · claude-code · moonshot/kimi-k3260kreview · task version 2 · codex · anthropic/claude-opus-5-5: 65k tokens per pass (5/6 passed)review · task version 2 · codex · anthropic/claude-opus-5-565kreview · task version 2 · codex · moonshot/kimi-k3: 58k tokens per pass (5/6 passed)review · task version 2 · codex · moonshot/kimi-k358kreview · task version 2 · opencode · anthropic/claude-opus-5-5: 74k tokens per pass (5/6 passed)review · task version 2 · opencode · anthropic/claude-opus-5-574kreview · task version 2 · opencode · fireworks/kimi-k3: 98k tokens per pass (5/6 passed)review · task version 2 · opencode · fireworks/kimi-k398kreview · task version 2 · opencode · fireworks/qwen3.8-max: 118k tokens per pass (5/6 passed)review · task version 2 · opencode · fireworks/qwen3.8-max118kreview · task version 2 · opencode · moonshot/kimi-k3: 95k tokens per pass (5/6 passed)review · task version 2 · opencode · moonshot/kimi-k395kreview · task version 2 · pi · anthropic/claude-opus-5-5: 28k tokens per pass (5/6 passed)review · task version 2 · pi · anthropic/claude-opus-5-528kreview · task version 2 · codex · fireworks/kimi-k3: 80k tokens per pass (8/10 passed)review · task version 2 · codex · fireworks/kimi-k380kreview · task version 2 · pi · fireworks/qwen3.8-max: 255k tokens per pass (8/10 passed)review · task version 2 · pi · fireworks/qwen3.8-max255kreview · task version 2 · claude-code · fireworks/qwen3.8-max: 357k tokens per pass (7/10 passed)review · task version 2 · claude-code · fireworks/qwen3.8-max357kreview · task version 2 · opencode · baseten/glm-5.3-flash: 142k tokens per pass (7/10 passed)review · task version 2 · opencode · baseten/glm-5.3-flash142kreview · task version 2 · claude-code · zai/glm-5.2: 445k tokens per pass (4/6 passed)review · task version 2 · claude-code · zai/glm-5.2445kreview · task version 2 · codex · anthropic/claude-sonnet-5-5: 70k tokens per pass (4/6 passed)review · task version 2 · codex · anthropic/claude-sonnet-5-570kreview · task version 2 · codex · baseten/glm-5.3-flash: 65k tokens per pass (4/6 passed)review · task version 2 · codex · baseten/glm-5.3-flash65kreview · task version 2 · codex · fireworks/qwen3.8-max: 170k tokens per pass (4/6 passed)review · task version 2 · codex · fireworks/qwen3.8-max170kreview · task version 2 · codex · together/deepseek-v4.1-flash: 101k tokens per pass (4/6 passed)review · task version 2 · codex · together/deepseek-v4.1-flash101kreview · task version 2 · opencode · zai/glm-5.2: 190k tokens per pass (4/6 passed)review · task version 2 · opencode · zai/glm-5.2190kreview · task version 2 · pi · zai/glm-5.2: 132k tokens per pass (4/6 passed)review · task version 2 · pi · zai/glm-5.2132kreview · task version 2 · codex · zai/glm-5.2: 397k tokens per pass (5/8 passed)review · task version 2 · codex · zai/glm-5.2397kreview · task version 2 · pi · baseten/glm-5.3-flash: 60k tokens per pass (6/10 passed)review · task version 2 · pi · baseten/glm-5.3-flash60kreview · task version 3 · codex · openai/gpt-6.1-sol @medium: 107k tokens per pass (23/23 passed)review · task version 3 · codex · openai/gpt-6.1-sol @medium107kreview · task version 3 · codex · openai/gpt-6.1-sol @high: 108k tokens per pass (21/21 passed)review · task version 3 · codex · openai/gpt-6.1-sol @high108kreview · task version 3 · claude-code · anthropic/claude-opus-5-5: 789k tokens per pass (12/12 passed)review · task version 3 · claude-code · anthropic/claude-opus-5-5789kreview · task version 3 · codex · openai/gpt-6.1-sol: 96k tokens per pass (12/12 passed)review · task version 3 · codex · openai/gpt-6.1-sol96kreview · task version 3 · claude-code · anthropic/claude-opus-5-5 @xhigh: 1.26M tokens per pass (10/10 passed)review · task version 3 · claude-code · anthropic/claude-opus-5-5 @xhigh1.26Mreview · task version 3 · codex · openai/gpt-6.1-sol @xhigh: 114k tokens per pass (10/10 passed)review · task version 3 · codex · openai/gpt-6.1-sol @xhigh114kreview · task version 3 · pi · google/gemini-3.8-flash: 397k tokens per pass (10/10 passed)review · task version 3 · pi · google/gemini-3.8-flash397kreview · task version 3 · antigravity · google/gemini-3.8-flash @high: 827k tokens per pass (9/9 passed)review · task version 3 · antigravity · google/gemini-3.8-flash @high827kreview · task version 3 · opencode · openai/gpt-6.1-sol: 129k tokens per pass (7/7 passed)review · task version 3 · opencode · openai/gpt-6.1-sol129kreview · task version 3 · antigravity · google/gemini-3.1-pro @high: 288k tokens per pass (6/6 passed)review · task version 3 · antigravity · google/gemini-3.1-pro @high288kreview · task version 3 · claude-code · anthropic/claude-sonnet-5-5: 767k tokens per pass (6/6 passed)review · task version 3 · claude-code · anthropic/claude-sonnet-5-5767kreview · task version 3 · opencode · together/deepseek-v4.1-flash: 82k tokens per pass (5/5 passed)review · task version 3 · opencode · together/deepseek-v4.1-flash82kreview · task version 3 · antigravity · google/gemini-4-argon @high: 249k tokens per pass (3/3 passed)review · task version 3 · antigravity · google/gemini-4-argon @high249kreview · task version 3 · claude-code · google/gemini-3.8-flash: 802k tokens per pass (3/3 passed)review · task version 3 · claude-code · google/gemini-3.8-flash802kreview · task version 3 · claude-code · google/gemini-3.8-flash @medium: 481k tokens per pass (3/3 passed)review · task version 3 · claude-code · google/gemini-3.8-flash @medium481kreview · task version 3 · claude-code · openai/gpt-6.1-sol: 194k tokens per pass (3/3 passed)review · task version 3 · claude-code · openai/gpt-6.1-sol194kreview · task version 3 · codex · google/gemini-3.8-flash: 501k tokens per pass (3/3 passed)review · task version 3 · codex · google/gemini-3.8-flash501kreview · task version 3 · codex · google/gemini-3.8-flash @low: 244k tokens per pass (3/3 passed)review · task version 3 · codex · google/gemini-3.8-flash @low244kreview · task version 3 · opencode · anthropic/claude-sonnet-5-5: 77k tokens per pass (3/3 passed)review · task version 3 · opencode · anthropic/claude-sonnet-5-577kreview · task version 3 · opencode · google/gemini-3.8-flash: 450k tokens per pass (3/3 passed)review · task version 3 · opencode · google/gemini-3.8-flash450kreview · task version 3 · opencode · google/gemini-3.8-flash @high: 565k tokens per pass (3/3 passed)review · task version 3 · opencode · google/gemini-3.8-flash @high565kreview · task version 3 · pi · anthropic/claude-sonnet-5-5: 26k tokens per pass (3/3 passed)review · task version 3 · pi · anthropic/claude-sonnet-5-526kreview · task version 3 · pi · fireworks/kimi-k3: 67k tokens per pass (3/3 passed)review · task version 3 · pi · fireworks/kimi-k367kreview · task version 3 · pi · google/gemini-3.8-flash @high: 693k tokens per pass (3/3 passed)review · task version 3 · pi · google/gemini-3.8-flash @high693kreview · task version 3 · pi · moonshot/kimi-k3: 59k tokens per pass (3/3 passed)review · task version 3 · pi · moonshot/kimi-k359kreview · task version 3 · pi · openai/gpt-6.1-sol: 39k tokens per pass (3/3 passed)review · task version 3 · pi · openai/gpt-6.1-sol39kreview · task version 3 · pi · together/deepseek-v4.1-flash: 129k tokens per pass (11/12 passed)review · task version 3 · pi · together/deepseek-v4.1-flash129kreview · task version 3 · claude-code · together/deepseek-v4.1-flash: 351k tokens per pass (7/8 passed)review · task version 3 · claude-code · together/deepseek-v4.1-flash351kreview · task version 3 · codex · fireworks/kimi-k3: 109k tokens per pass (5/7 passed)review · task version 3 · codex · fireworks/kimi-k3109kreview · task version 3 · pi · fireworks/qwen3.8-max: 382k tokens per pass (5/7 passed)review · task version 3 · pi · fireworks/qwen3.8-max382kreview · task version 3 · claude-code · baseten/glm-5.3-flash: 737k tokens per pass (2/3 passed)review · task version 3 · claude-code · baseten/glm-5.3-flash737kreview · task version 3 · claude-code · fireworks/kimi-k3: 481k tokens per pass (2/3 passed)review · task version 3 · claude-code · fireworks/kimi-k3481kreview · task version 3 · claude-code · google/gemini-3.8-flash @high: 779k tokens per pass (2/3 passed)review · task version 3 · claude-code · google/gemini-3.8-flash @high779kreview · task version 3 · claude-code · moonshot/kimi-k3: 385k tokens per pass (2/3 passed)review · task version 3 · claude-code · moonshot/kimi-k3385kreview · task version 3 · codex · anthropic/claude-opus-5-5: 86k tokens per pass (2/3 passed)review · task version 3 · codex · anthropic/claude-opus-5-586kreview · task version 3 · codex · baseten/glm-5.3-flash: 71k tokens per pass (2/3 passed)review · task version 3 · codex · baseten/glm-5.3-flash71kreview · task version 3 · codex · moonshot/kimi-k3: 92k tokens per pass (2/3 passed)review · task version 3 · codex · moonshot/kimi-k392kreview · task version 3 · opencode · anthropic/claude-opus-5-5: 100k tokens per pass (2/3 passed)review · task version 3 · opencode · anthropic/claude-opus-5-5100kreview · task version 3 · opencode · fireworks/kimi-k3: 156k tokens per pass (2/3 passed)review · task version 3 · opencode · fireworks/kimi-k3156kreview · task version 3 · opencode · fireworks/qwen3.8-max: 192k tokens per pass (2/3 passed)review · task version 3 · opencode · fireworks/qwen3.8-max192kreview · task version 3 · opencode · google/gemini-3.8-flash @low: 144k tokens per pass (2/3 passed)review · task version 3 · opencode · google/gemini-3.8-flash @low144kreview · task version 3 · opencode · moonshot/kimi-k3: 150k tokens per pass (2/3 passed)review · task version 3 · opencode · moonshot/kimi-k3150kreview · task version 3 · pi · anthropic/claude-opus-5-5: 38k tokens per pass (2/3 passed)review · task version 3 · pi · anthropic/claude-opus-5-538kreview · task version 3 · claude-code · fireworks/qwen3.8-max: 467k tokens per pass (6/10 passed)review · task version 3 · claude-code · fireworks/qwen3.8-max467kreview · task version 3 · opencode · baseten/glm-5.3-flash: 212k tokens per pass (4/7 passed)review · task version 3 · opencode · baseten/glm-5.3-flash212kreview · task version 3 · codex · zai/glm-5.2: 772k tokens per pass (4/8 passed)review · task version 3 · codex · zai/glm-5.2772kreview · task version 3 · claude-code · fireworks/qwen3.8-27b: 323k tokens per pass (1/2 passed)review · task version 3 · claude-code · fireworks/qwen3.8-27b323kreview · task version 3 · pi · baseten/glm-5.3-flash: 103k tokens per pass (3/7 passed)review · task version 3 · pi · baseten/glm-5.3-flash103kreview · task version 3 · claude-code · zai/glm-5.2: 1.52M tokens per pass (2/6 passed)review · task version 3 · claude-code · zai/glm-5.21.52Mreview · task version 3 · codex · together/deepseek-v4.1-flash: 283k tokens per pass (2/6 passed)review · task version 3 · codex · together/deepseek-v4.1-flash283kreview · task version 3 · pi · zai/glm-5.2: 636k tokens per pass (2/6 passed)review · task version 3 · pi · zai/glm-5.2636kreview · task version 3 · claude-code · google/gemini-3.8-flash @low: 358k tokens per pass (1/3 passed)review · task version 3 · claude-code · google/gemini-3.8-flash @low358kreview · task version 3 · codex · anthropic/claude-sonnet-5-5: 149k tokens per pass (1/3 passed)review · task version 3 · codex · anthropic/claude-sonnet-5-5149kreview · task version 3 · codex · fireworks/qwen3.8-max: 510k tokens per pass (1/3 passed)review · task version 3 · codex · fireworks/qwen3.8-max510kreview · task version 3 · opencode · zai/glm-5.2: 1.09M tokens per pass (1/6 passed)review · task version 3 · opencode · zai/glm-5.21.09M