Broad-Domain Factual and Professional Reasoning on MMLU PRO
0.711Accuracy (ACC)Raw CoT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Raw CoTLLM=DEEPSEEK-R1-0528-QWEN3-8B2026.06 | 0.711 | 1,789 | |
| CLSRLLM=DEEPSEEK-R1-0528-QWEN3-8B2026.06 | 0.703 | 356 | |
| SoTLLM=DEEPSEEK-R1-0528-QWEN3-8B2026.06 | 0.692 | 458 | |
| CoDLLM=DEEPSEEK-R1-0528-QWEN3-8B2026.06 | 0.683 | 582 | |
| CLSRLLM=QWEN3-32B2026.06 | 0.681 | 118 | |
| Raw CoTLLM=QWEN3-32B2026.06 | 0.675 | 405 | |
| CCoTLLM=QWEN3-32B2026.06 | 0.651 | 192 | |
| CCoTLLM=DEEPSEEK-R1-0528-QWEN3-8B2026.06 | 0.645 | 516 | |
| SoTLLM=QWEN3-32B2026.06 | 0.643 | 134 | |
| CoDLLM=QWEN3-32B2026.06 | 0.642 | 318 | |
| CLSRLLM=QWEN3-8B2026.06 | 0.604 | 96 | |
| Raw CoTLLM=QWEN3-8B2026.06 | 0.602 | 276 | |
| SoTLLM=QWEN3-8B2026.06 | 0.585 | 118 | |
| CoDLLM=QWEN3-8B2026.06 | 0.556 | 125 | |
| CCoTLLM=QWEN3-8B2026.06 | 0.502 | 135 | |
| CLSRLLM=LLaMA3-8B2026.06 | 0.389 | 320 | |
| Raw CoTLLM=LLaMA3-8B2026.06 | 0.386 | 960 | |
| CCoTLLM=LLaMA3-8B2026.06 | 0.372 | 370 | |
| SoTLLM=LLaMA3-8B2026.06 | 0.364 | 340 | |
| CoDLLM=LLaMA3-8B2026.06 | 0.358 | 352 |