Mathematical Reasoning on MATH 500 (Accuracy and Delta Change)
98.8AccuracyDeepSeek-R1-0528
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepSeek-R1-0528Thinking Mode=Thinking2026.04 | 98.8 | 7.2 | |
| Qwen3-30B-A3B-ThinkingThinking Mode=Thinking2026.04 | 97.5 | 10.5 | |
| Qwen3-8BThinking Mode=Thinking2026.04 | 96.4 | 11.1 | |
| DeepSeek-V3.1Thinking Mode=No Thinking2026.04 | 92.2 | — | |
| Qwen3-30B-A3B-InstructThinking Mode=No Thinking2026.04 | 88.2 | — | |
| Qwen3-8BThinking Mode=No Thinking2026.04 | 86.8 | — |