Board · task version 3

docs

Each board ranks harness × model combinations on one suite by pass rate with a 95% Wilson interval; rows are not comparable across suites or task versions.

3 tasks
  • docs-gym-baseline-keys
  • docs-retry-behavior-diff
  • docs-snapshot-new-command
docs · task version 3
#Liverymodel · harness · configpassed / n95% Wilsonmeantaskstokens / passcost / passexcluded
1google/gemini-4-argon @highantigravity · config 034ed3a13 / 3 44%–100%1.003114k$0.11 list1
2anthropic/claude-sonnet-5-5claude-code · config c32a950b3 / 3 44%–100%1.0031.14M$0.17 list0
3baseten/glm-5.3-flashclaude-code · config ecb67e4e3 / 3 44%–100%1.003257k$0.03 list0
4fireworks/kimi-k3claude-code · config ecb67e4e3 / 3 44%–100%1.003180k$0.13 list0
5fireworks/qwen3.8-maxclaude-code · config ecb67e4e3 / 3 44%–100%1.003230k$0.11 list0
6google/gemini-3.8-flashclaude-code · config ecb67e4e3 / 3 44%–100%1.003388k$0.07 list0
7google/gemini-3.8-flash @highclaude-code · config ecb67e4e3 / 3 44%–100%1.003457k$0.08 list0
8moonshot/kimi-k3claude-code · config ecb67e4e3 / 3 44%–100%1.003210k$0.12 list0
9together/deepseek-v4.1-flashclaude-code · config ecb67e4e3 / 3 44%–100%1.003262k$0.08 list0
10zai/glm-5.2claude-code · config ecb67e4e3 / 3 44%–100%1.003198k$0.07 list0
11anthropic/claude-opus-5-5codex · config 3437a9e63 / 3 44%–100%1.00346k$0.20 list0
12anthropic/claude-sonnet-5-5codex · config 3437a9e63 / 3 44%–100%1.00334k$0.07 list0
13baseten/glm-5.3-flashcodex · config 3437a9e63 / 3 44%–100%1.00349k$0.0070 list0
14fireworks/kimi-k3codex · config 3437a9e63 / 3 44%–100%1.00336k$0.05 list0
15google/gemini-3.8-flashcodex · config 3437a9e63 / 3 44%–100%1.003131k$0.09 list0
16moonshot/kimi-k3codex · config 3437a9e63 / 3 44%–100%1.00354k$0.05 list0
17openai/gpt-6.1-solcodex · config b23964033 / 3 44%–100%1.00376k$0.03 list0
18openai/gpt-6.1-sol @highcodex · config b23964033 / 3 44%–100%1.00389k$0.04 list0
19anthropic/claude-opus-5-5opencode · config 357da8913 / 3 44%–100%1.00351k$0.10 list0
20anthropic/claude-sonnet-5-5opencode · config 357da8913 / 3 44%–100%1.00357k$0.05 list0
21fireworks/kimi-k3opencode · config 357da8913 / 3 44%–100%1.00372k$0.11 list0
22fireworks/qwen3.8-maxopencode · config 357da8913 / 3 44%–100%1.00377k$0.06 list0
23moonshot/kimi-k3opencode · config 357da8913 / 3 44%–100%1.00372k$0.08 list0
24anthropic/claude-opus-5-5pi · config da41c66e3 / 3 44%–100%1.00321k$0.11 list0
25anthropic/claude-sonnet-5-5pi · config da41c66e3 / 3 44%–100%1.00317k$0.04 list0
26baseten/glm-5.3-flashpi · config da41c66e3 / 3 44%–100%1.00321k$0.0036 list0
27fireworks/kimi-k3pi · config da41c66e3 / 3 44%–100%1.00328k$0.05 list0
28fireworks/qwen3.8-maxpi · config da41c66e3 / 3 44%–100%1.00356k$0.06 list0
29zai/glm-5.2pi · config da41c66e3 / 3 44%–100%1.00333k$0.03 list0
30google/gemini-3.8-flash @highantigravity · config 034ed3a12 / 3 21%–94%0.673423k$0.19 list0
31google/gemini-3.8-flash @lowclaude-code · config ecb67e4e2 / 3 21%–94%0.963177k$0.14 list0
32google/gemini-3.8-flash @mediumclaude-code · config ecb67e4e2 / 3 21%–94%0.963230k$0.18 list0
33openai/gpt-6.1-solclaude-code · config ecb67e4e2 / 3 21%–94%0.973257k$0.11 list0
34fireworks/qwen3.8-maxcodex · config 3437a9e62 / 3 21%–94%0.673166k$0.15 list0
35google/gemini-3.8-flash @highcodex · config 3437a9e62 / 3 21%–94%0.973380k$0.18 list0
36google/gemini-3.8-flash @lowcodex · config 3437a9e62 / 3 21%–94%0.963121k$0.10 list0
37together/deepseek-v4.1-flashcodex · config 3437a9e62 / 3 21%–94%0.67376k$0.02 list0
38zai/glm-5.2codex · config 3437a9e62 / 3 21%–94%0.96383k$0.06 list0
39openai/gpt-6.1-sol @mediumcodex · config b23964032 / 3 21%–94%0.933124k$0.07 list0
40baseten/glm-5.3-flashopencode · config 357da8912 / 3 21%–94%0.673100k$0.01 list0
41google/gemini-3.8-flashopencode · config 357da8912 / 3 21%–94%0.673180k$0.14 list0
42google/gemini-3.8-flash @highopencode · config 357da8912 / 3 21%–94%0.963382k$0.21 list0
43openai/gpt-6.1-solopencode · config 357da8912 / 3 21%–94%0.673117k$0.09 list0
44together/deepseek-v4.1-flashopencode · config 357da8912 / 3 21%–94%0.67372k$0.02 list0
45zai/glm-5.2opencode · config 357da8912 / 3 21%–94%0.67396k$0.06 list0
46google/gemini-3.8-flashpi · config da41c66e2 / 3 21%–94%0.673112k$0.10 list0
47google/gemini-3.8-flash @highpi · config da41c66e2 / 3 21%–94%0.963120k$0.11 list0
48google/gemini-3.8-flash @lowpi · config da41c66e2 / 3 21%–94%0.96354k$0.04 list0
49moonshot/kimi-k3pi · config da41c66e2 / 3 21%–94%0.67350k$0.09 list0
50openai/gpt-6.1-solpi · config da41c66e2 / 3 21%–94%0.67334k$0.03 list0
51together/deepseek-v4.1-flashpi · config da41c66e2 / 3 21%–94%0.67334k$0.01 list0
52anthropic/claude-opus-5-5claude-code · config c32a950b1 / 3 6%–79%0.9333.43M$1.08 list0
53google/gemini-3.8-flash @lowopencode · config 357da8911 / 3 6%–79%0.933324k$0.25 list0
54google/gemini-3.1-pro @highantigravity · config 034ed3a10 / 3 0%–56%1no passesunknown0
list = published-rate · invoice = provider-invoice-reconciled