| Created | State | Score | Judge | Model | Tokens | Reason | Record |
|---|
| 2026-07-08 | promoted | 36.4% 16/44 | — | gpt-5.5 | 0 | | record |
| 2026-06-24 | yanked | 97.7% 43/44 | — | gpt-5.5 | 0 | Finding-fed task mode: the scaffold injected each audit finding (root cause + recommended fix) and the target file, so this run measures 'implement a described fix', not the detect+patch task the 41.5% Patch SOTA measures. Not comparable to the SOTA; removed from the board pending a blind-audit re-run. See docs/PHASE2_GAMEABILITY.md section 4b. | record |
| 2026-06-21 | promoted | 54.7% 64/117 | gpt-5.5 (high) | claude-opus-4-8+gpt-5.5 consensus | 0 | | record |
| 2026-06-21 | promoted | 35.0% 41/117 | gpt-5.5 (high) | composer-2.5 | 0 | | record |
| 2026-06-21 | promoted | 35.9% 42/117 | gpt-5.5 (high) | claude-opus-4-8 | 0 | | record |
| 2026-06-20 | promoted | 71.8% 84/117 | gpt-5.5 (high) | gpt-5.5 | 0 | | record |
| 2026-06-20 | promoted | 59.0% 69/117 | gpt-5.5 (high) | gpt-5.5 | 0 | | record |
| 2026-06-20 | promoted | 4.3% 5/117 | gpt-5.5 (high) | llama-3-3-70b | 0 | | record |
| 2026-06-20 | promoted | 9.4% 11/117 | gpt-5.5 (high) | x-ai-grok-4-3 | 0 | | record |
| 2026-06-20 | promoted | 28.2% 33/117 | gpt-5.5 (high) | claude-opus-4-8 | 0 | | record |
| 2026-06-20 | promoted | 13.7% 16/117 | gpt-5.5 (high) | google-gemini-3-1-pro-preview | 0 | | record |
| 2026-06-20 | promoted | 26.5% 31/117 | gpt-5.5 (high) | google-gemini-3-5-flash | 0 | | record |
| 2026-06-20 | promoted | 26.5% 31/117 | gpt-5.5 (high) | moonshotai-kimi-k2-7-code | 0 | | record |
| 2026-06-20 | promoted | 43.6% 51/117 | gpt-5.5 (high) | gpt-5.4 | 0 | | record |
| 2026-06-19 | promoted | 17.9% 21/117 | gpt-5.5 (high) | minimax-m3 | 0 | | record |
| 2026-06-19 | promoted | 61.5% 72/117 | gpt-5 (high) | gpt-5.5 | 0 | | record |
| 2026-06-19 | promoted | 27.4% 32/117 | gpt-5 (high) | zai-org-glm-5-2 | 0 | | record |
| 2026-06-19 | promoted | 18.8% 22/117 | gpt-5 (high) | qwen/qwen3.7-max | 0 | | record |
| 2026-06-18 | promoted | 16.2% 19/117 | gpt-5 (high) | deepseek-v4-pro | 0 | | record |
| 2026-06-18 | promoted | 43.6% 51/117 | gpt-5 (high) | claude-opus-4-8+gpt-5.5 consensus | 0 | | record |