STEM Reasoning on MMLU-Pro
90.8Pass@1 AccuracyQwen3-30B-A3B (Thinking)
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-30B-A3B (Thinking)Sampling Strategy=4-sample mean2026.04 | 90.8 | |
| Gemini 2.5 FlashSampling Strategy=4-sample mean2026.04 | 90.44 | |
| GPT-OSS-120BSampling Strategy=4-sample mean2026.04 | 90.11 | |
| GPT-5 MiniSampling Strategy=4-sample mean2026.04 | 89.64 | |
| Aryabhata 2Sampling Strategy=4-sample mean2026.04 | 88.49 | |
| GPT-OSS-20BSampling Strategy=4-sample mean2026.04 | 85.42 | |
| Nemotron 3 Nano 30B A3BSampling Strategy=4-sample mean2026.04 | 84.33 | |
| GPT-5 NanoSampling Strategy=4-sample mean2026.04 | 80.62 |