FreshDirect
Rebuilt by Claude Opus 5.5: 292 of 990 hidden tests passed (29.5%)
reference.local/
Referencerebuild.local/
Claude Opus 5.5| Model | Passed | % | Screenshot |
|---|---|---|---|
| 292 | 29.5 | ||
| 246 | 24.8 | ||
| 221 | 22.3 | ||
| 184 | 18.6 | ||
| 153 | 15.5 | ||
| 133 | 13.4 | ||
| 130 | 13.1 | ||
| 127 | 12.8 |