STEM Knowledge on GPQA Diamond
90.7Avg@16Gemini-3-Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-FlashTier=High2026.06 | 90.7 | |
| Seed2.0 Lite2026.06 | 85.1 | |
| GPT-5-miniTier=High2026.06 | 82.1 | |
| Seed2.0 Mini2026.06 | 79 | |
| LongCat-Flash Exp-ChatEvaluation Mode=Chat2025.12 | 75.6 | |
| DeepSeek V3.2Evaluation Mode=Chat2025.12 | 75.3 | |
| GLM 4.6Evaluation Mode=Chat2025.12 | 74.8 | |
| LongCat-Flash ChatEvaluation Mode=Chat2025.12 | 73.2 |