GeneBench-Pro Externally Reviewed
Science · 2026-06-30
GeneBench-Pro externally reviewed stratum reports pass rates on the 82 externally reviewed problems in the 129-problem GeneBench-Pro suite.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.6 Sol Pro | 28.5 |
| GPT-5.6 Terra Pro | 27.3 |
| GPT-5.6 Sol | 25.7 |
| GPT-5.6 Terra | 20.2 |
| GPT-5.6 Luna Pro | 20.2 |
| GPT-5.5 Pro | 18.0 |
| GPT-5.4 Pro | 15.4 |
| GPT-5.6 Luna | 13.9 |
| Opus 4.8 | 13.4 |
| GPT-5.5 | 9.3 |
| GPT-5.2 Pro | 7.3 |
| GPT-5.4 | 6.9 |
| Gemini 3.5 Flash | 5.4 |
| Kimi K2.6 | 4.0 |
| Qwen3.7-Max | 3.9 |
| GLM-5.2 | 3.9 |
| GPT-5.2 | 3.7 |
| Gemini 3.1 Pro Preview | 2.9 |
| DeepSeek-V4-Pro | 2.4 |
| MiMo-V2.5-Pro | 2.1 |
| Qwen3.7-Plus | 2.0 |
| DeepSeek-V4-Flash | 2.0 |
| Kimi K2.7 Code | 1.8 |
| GLM-5.1 | 1.2 |
| MiMo-V2.5 | 1.1 |
| Hy3 preview | 0.9 |
| MiniMax M3 | 0.6 |
| Grok 4.3 | 0.6 |
| MiniMax M2.7 | 0.5 |