Multi-step Reasoning on GSM-Hard
71.5AccuracyeMoT
Evaluation Results
| Method | Links | |
|---|---|---|
| eMoTModels=Qwen-32B2026.06 | 71.5 | |
| BoTModels=GPT-42026.06 | 62 | |
| PaLModels=Codex (175B)2026.06 | 61.2 | |
| ToTModels=GPT-42026.06 | 24.4 | |
| Qwen-32B (Direct)Models=Qwen-32B2026.06 | 15.9 |
| Method | Links | |
|---|---|---|
| eMoTModels=Qwen-32B2026.06 | 71.5 | |
| BoTModels=GPT-42026.06 | 62 | |
| PaLModels=Codex (175B)2026.06 | 61.2 | |
| ToTModels=GPT-42026.06 | 24.4 | |
| Qwen-32B (Direct)Models=Qwen-32B2026.06 | 15.9 |