Instruction Following on COLLIE
97.4ScoreGPT-5-mini
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5-miniTier=High2026.06 | 97.4 | |
| GPT-5.2 High2026.06 | 96.9 | |
| Gemini-3-FlashTier=High2026.06 | 96.5 | |
| Gemini-3-pro High2026.06 | 95 | |
| Seed2.0 Lite2026.06 | 94 | |
| Seed2.0 Pro2026.06 | 93.9 | |
| Seed2.0 Mini2026.06 | 91.2 | |
| Claude-Opus-4.52026.06 | 79.8 | |
| Claude-Sonnet-4.52026.06 | 77.3 |