Visual Mathematical Reasoning on FlowVerse Vision Primary RP+EI+OQ
69.4CoT-E ScoreMathFlow
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MathFlowBackbone=GPT-52025.03 | 69.4 | 66.2 | |
| GPT-52025.03 | 60.3 | 57.5 | |
| InternVL2.5-78B2025.03 | 59.6 | 57.7 | |
| Gemini-2.5-pro2025.03 | 57 | 54.6 | |
| GPT-4o2025.03 | 54.1 | 51 | |
| MathFlowBackbone=Qwen2.5-VL-7B2025.03 | 52 | 51.5 | |
| GPT-4o-mini2025.03 | 49.7 | 47.6 | |
| Qwen2.5-VL-7B2025.03 | 48.1 | 45.3 | |
| Claude-3.5-Sonnet2025.03 | 48.1 | 45.2 | |
| Qwen2-VL-72B2025.03 | 47.6 | 37 | |
| SVE-Math-Qwen2.5-7B2025.03 | 45.8 | 42 | |
| GPT-4V2025.03 | 42 | 36.7 | |
| MultiMath-7B2025.03 | 39.7 | 38.8 | |
| MathLLaVA-13B2025.03 | 37 | 34.2 | |
| InternVL2.5-8B2025.03 | 35.8 | 33.9 | |
| Gemini-Pro2025.03 | 33.8 | 24.9 | |
| VLM-R1-7B†2025.03 | 30.6 | 26.1 | |
| Qwen-VL-Plus2025.03 | 30 | 22.2 | |
| Qwen-VL-Max2025.03 | 29.6 | 27.8 | |
| InfiMM-Math-7B2025.03 | 26.1 | 23.2 |