The visible test checks for a specific bug. A naive fix makes the visible test pass but introduces a subtle issue in a related code path. The hidden spec catches the wrong fix. Read the entire src/ to find the real fix.
| # | Model | Status | Duration | Tokens | Patch | Failure |
|---|---|---|---|---|---|---|
| 1 | defaultinternal | pass | 1m 51s | 56.1k | yes | - |
| 2 | MiniMax-M3minimax | pass | 1m 56s | 80.5k | yes | - |
| 3 | kimi-k3kimi | pass | 3m 3s | 74.9k | yes | - |
| 4 | deepseek-v4-pro+qwen3.8-max-previewconsensus | pass | 5m 11s | 202.6k | - | - |
| 5 | deepseek-v4-prodeepseek | pass | 6m 17s | 125.5k | yes | - |
| 6 | qwen3.8-max-previewqwen | done | 5m 11s | 102.2k | yes | - |