Reasoning on HLE no tools
37.5ScoreGemini-3.0 Pro
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini-3.0 ProVariant=Pro2026.01 | 37.5 | |
| GPT-5 HighVariant=High2026.01 | 26.3 | |
| DeepSeek-V3.2 ThinkingThinking Mode=true2026.01 | 25.1 | |
| Kimi-K2 ThinkingThinking Mode=true2026.01 | 23.9 | |
| MiMo-V2-FlashVariant=Flash2026.01 | 22.1 | |
| SuperPrecision=BF162026.07 | 18.5 | |
| SuperPrecision=NVFP42026.07 | 17.9 | |
| Puzzle-75B-A9BPrecision=BF162026.07 | 16.5 | |
| Puzzle-75B-A9BPrecision=NVFP42026.07 | 15.7 | |
| Claude Sonnet 4.5Variant=Sonnet 4.52026.01 | 13.7 |