Mathematical Reasoning on MathVerse Vision Intensive (testmini)
53.7CoT-EMathFlow
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MathFlowPerception Model=MathFlow-P-7B, Inference Model=GPT-4V2025.03 | 53.7 | 40.3 | |
| GPT-4V2025.03 | 51.4 | 34.9 | |
| MathFlowPerception Model=MathFlow-P-7B, Inference Model=Qwen2-VL-72B2025.03 | 40.5 | 30.3 | |
| Qwen2-VL-72B2025.03 | 38.7 | 26.4 | |
| InfiMM-Math2025.03 | 38.1 | — | |
| Qwen-VL-Max2025.03 | 33.6 | 24.1 | |
| Gemini-Pro2025.03 | 32 | 23 | |
| MAVIS-7B2025.03 | 27.4 | — | |
| SPHINX-MoE2025.03 | 21.1 | 14.8 | |
| InternLM-XC22025.03 | 20.1 | 15.7 | |
| Qwen-VL-Plus2025.03 | 18.5 | 9 |