Expert-level Reasoning on HLE
26.5Accuracy (HLE)Gemma 4 31B
Evaluation results
| Method | Links | |
|---|---|---|
| Gemma 4 31BModel size=31B, Thinking mode=true, Search=true2026.07 | 26.5 | |
| Gemma 4 31BModel size=31B, Thinking mode=true, Search=false2026.07 | 19.5 | |
| Gemma 4 26B-A4BModel size=26B-A4B, Thinking mode=true, Search=true2026.07 | 17.2 | |
| Gemma 4 26B-A4BModel size=26B-A4B, Thinking mode=true, Search=false2026.07 | 8.7 | |
| Gemma 4 12BModel size=12B, Thinking mode=true, Search=false2026.07 | 5.2 |