Scientific Verification on FMT
70.5Exact Match (EM)PRISM-MCTS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PRISM-MCTSModel=GPT-4.1-mini2026.04 | 70.5 | 70.07 | |
| MCTS-RAGModel=Qwen3-30B-A3B-Instruct-25072026.04 | 70.5 | 69.99 | |
| MCTS-RAGModel=GPT-4.1-mini2026.04 | 69.5 | 68.98 | |
| ReST-MCTS*Model=GPT-4.1-mini2026.04 | 69 | 68.91 | |
| Zero-shotModel=GPT-4.1-mini2026.04 | 68.34 | 68.75 | |
| Search-O1Model=Qwen3-30B-A3B-Instruct-25072026.04 | 68 | 67.76 | |
| PRISM-MCTSModel=Qwen3-30B-A3B-Instruct-25072026.04 | 68 | 67.41 | |
| Search-O1Model=GPT-4.1-mini2026.04 | 66 | 65.03 | |
| ReActModel=Qwen3-30B-A3B-Instruct-25072026.04 | 65.5 | 65.43 | |
| ReActModel=GPT-4.1-mini2026.04 | 65 | 67.61 | |
| ReST-MCTS*Model=Qwen3-30B-A3B-Instruct-25072026.04 | 64 | 63.43 | |
| Zero-shotModel=Qwen3-30B-A3B-Instruct-25072026.04 | 62.72 | 62.85 |