Board · task version 2

review

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.

6 tasks
  • review-clean-retention-plan
  • review-corrupt-refresh-checkpoint
  • review-publish-guard-order
  • review-rescue-local-tip
  • review-retry-empty-page
  • review-stale-lock-recovery
review · task version 2
#Liverymodel · harness · configpassed / n95% Wilsonmeantaskstokens / passcost / passexcluded
1openai/gpt-6.1-sol @mediumcodex · config b239640323 / 23 86%–100%0.995103k$0.04 list0
2google/gemini-3.8-flash @highantigravity · config 034ed3a118 / 18 82%–100%0.966589k$0.29 list0
3openai/gpt-6.1-sol @highcodex · config b239640314 / 14 78%–100%0.976101k$0.05 list2
4anthropic/claude-opus-5-5claude-code · config c32a950b12 / 12 76%–100%0.976757k$0.30 list0
5openai/gpt-6.1-solcodex · config b239640312 / 12 76%–100%0.98694k$0.04 list0
6openai/gpt-6.1-sol @xhighcodex · config b239640310 / 10 72%–100%1.002114k$0.06 list0
7openai/gpt-6.1-solopencode · config 357da89110 / 10 72%–100%0.966109k$0.08 list0
8together/deepseek-v4.1-flashpi · config da41c66e10 / 10 72%–100%0.98689k$0.03 list0
9together/deepseek-v4.1-flashopencode · config 357da8919 / 9 70%–100%0.93660k$0.02 list0
10google/gemini-3.1-pro @highantigravity · config 034ed3a16 / 6 61%–100%0.936278kunknown0
11google/gemini-4-argon @highantigravity · config 034ed3a16 / 6 61%–100%0.976176k$0.32 list0
12google/gemini-3.8-flashclaude-code · config ecb67e4e6 / 6 61%–100%0.976610k$0.23 list0
13openai/gpt-6.1-solclaude-code · config ecb67e4e6 / 6 61%–100%0.936169k$0.08 list0
14google/gemini-3.8-flashcodex · config 3437a9e66 / 6 61%–100%0.976423k$0.23 list0
15anthropic/claude-sonnet-5-5opencode · config 357da8916 / 6 61%–100%0.93667k$0.06 list0
16google/gemini-3.8-flashopencode · config 357da8916 / 6 61%–100%0.936283k$0.15 list0
17anthropic/claude-sonnet-5-5pi · config da41c66e6 / 6 61%–100%0.97622k$0.05 list0
18fireworks/kimi-k3pi · config da41c66e6 / 6 61%–100%0.97642k$0.13 list0
19google/gemini-3.8-flashpi · config da41c66e6 / 6 61%–100%0.976273k$0.17 list0
20google/gemini-3.8-flash @highpi · config da41c66e6 / 6 61%–100%0.976440k$0.26 list0
21moonshot/kimi-k3pi · config da41c66e6 / 6 61%–100%0.97641k$0.12 list0
22openai/gpt-6.1-solpi · config da41c66e6 / 6 61%–100%0.97629k$0.03 list0
23fireworks/qwen3.8-27bclaude-code · config ecb67e4e3 / 3 44%–100%0.933192kunknown0
24google/gemini-3.8-flash @highclaude-code · config ecb67e4e3 / 3 44%–100%0.933523k$0.15 list0
25google/gemini-3.8-flash @lowclaude-code · config ecb67e4e3 / 3 44%–100%0.873118k$0.08 list0
26google/gemini-3.8-flash @mediumclaude-code · config ecb67e4e3 / 3 44%–100%0.933209k$0.14 list0
27google/gemini-3.8-flash @lowcodex · config 3437a9e63 / 3 44%–100%0.93382k$0.06 list0
28google/gemini-3.8-flash @highopencode · config 357da8913 / 3 44%–100%0.933336k$0.17 list0
29google/gemini-3.8-flash @lowopencode · config 357da8913 / 3 44%–100%0.93387k$0.07 list0
30together/deepseek-v4.1-flashclaude-code · config ecb67e4e7 / 8 53%–98%0.936268k$0.09 list0
31anthropic/claude-sonnet-5-5claude-code · config c32a950b5 / 6 44%–97%0.936930k$0.16 list0
32baseten/glm-5.3-flashclaude-code · config ecb67e4e5 / 6 44%–97%0.906435k$0.05 list0
33fireworks/kimi-k3claude-code · config ecb67e4e5 / 6 44%–97%0.906274k$0.17 list0
34moonshot/kimi-k3claude-code · config ecb67e4e5 / 6 44%–97%0.906260k$0.23 list0
35anthropic/claude-opus-5-5codex · config 3437a9e65 / 6 44%–97%0.87665k$0.29 list0
36moonshot/kimi-k3codex · config 3437a9e65 / 6 44%–97%0.90658k$0.12 list0
37anthropic/claude-opus-5-5opencode · config 357da8915 / 6 44%–97%0.87674k$0.13 list0
38fireworks/kimi-k3opencode · config 357da8915 / 6 44%–97%0.90698k$0.23 list1
39fireworks/qwen3.8-maxopencode · config 357da8915 / 6 44%–97%0.876118k$0.12 list0
40moonshot/kimi-k3opencode · config 357da8915 / 6 44%–97%0.90695k$0.18 list0
41anthropic/claude-opus-5-5pi · config da41c66e5 / 6 44%–97%0.87628k$0.14 list0
42fireworks/kimi-k3codex · config 3437a9e68 / 10 49%–94%0.86680k$0.15 list0
43fireworks/qwen3.8-maxpi · config da41c66e8 / 10 49%–94%0.866255k$0.21 list0
44fireworks/qwen3.8-maxclaude-code · config ecb67e4e7 / 10 40%–89%0.806357k$0.38 list3
45baseten/glm-5.3-flashopencode · config 357da8917 / 10 40%–89%0.846142k$0.02 list1
46zai/glm-5.2claude-code · config ecb67e4e4 / 6 30%–90%0.806445k$0.39 list0
47anthropic/claude-sonnet-5-5codex · config 3437a9e64 / 6 30%–90%0.90670k$0.15 list0
48baseten/glm-5.3-flashcodex · config 3437a9e64 / 6 30%–90%0.83665k$0.01 list0
49fireworks/qwen3.8-maxcodex · config 3437a9e64 / 6 30%–90%0.876170k$0.21 list3
50together/deepseek-v4.1-flashcodex · config 3437a9e64 / 6 30%–90%0.836101k$0.04 list0
51zai/glm-5.2opencode · config 357da8914 / 6 30%–90%0.806190k$0.16 list0
52zai/glm-5.2pi · config da41c66e4 / 6 30%–90%0.806132k$0.17 list0
53zai/glm-5.2codex · config 3437a9e65 / 8 31%–86%0.826397k$0.54 list0
54baseten/glm-5.3-flashpi · config da41c66e6 / 10 31%–83%0.84660k$0.01 list0
55google/gemini-3.8-flash @lowpi · config da41c66e— 0%–0%0not measured (0/0)unknown3
list = published-rate · invoice = provider-invoice-reconciled