Mathematical Reasoning on MathVision Mini
47.36ScoreQwen2.5-VL-32B-Instruct
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-VL-32B-Instruct2025.05 | 47.36 | |
| Individual Geometry2025.05 | 43.42 | |
| Individual Chart2025.05 | 43.42 | |
| OptMerge2025.05 | 43.42 | |
| Individual VQA2025.05 | 42.11 | |
| Individual Grounding2025.05 | 42.1 | |
| Innovator-VLvariant=8B-Thinking, parameters=8B2026.01 | 37.5 | |
| MiMo-VLtraining=7B-RL, parameters=7B2026.01 | 36.51 | |
| GPT-4oTool use=No2025.11 | 35.9 | |
| Individual OCR2025.05 | 34.21 | |
| MiMo-VLtraining=7B-SFT, parameters=7B2026.01 | 33.88 | |
| CodeV-7B-RLTool use=Yes, Training=RL2025.11 | 33.6 | |
| InternVL3.5parameters=8B2026.01 | 30.26 | |
| Qwen3-VLparameters=8B2026.01 | 29.61 | |
| Innovator-VLvariant=8B-Instruct, parameters=8B2026.01 | 28.95 | |
| LLaDA2.0-UniModel Category=Unified Models2026.04 | 26.7 | |
| LLaVA-OVversion=1.5, parameters=8B2026.01 | 26.64 | |
| InternVL3-8BTool use=No2025.11 | 26.3 | |
| DeepEyes-7BTool use=Yes2025.11 | 26.3 | |
| Pixel-Reasoner-7BTool use=Yes2025.11 | 26.3 | |
| ThinkLite-7BTool use=No2025.11 | 25.7 | |
| BAGELModel Category=Unified Models2026.04 | 24.1 | |
| Intern-S1variant=mini, parameters=9B2026.01 | 24.01 | |
| CodeV-7B-SFTTool use=Yes, Training=SFT2025.11 | 23.7 | |
| Qwen2.5-VL-7BModel Category=Specialist VLMs2026.04 | 22.4 | |
| InternVL-UModel Category=Unified Models2026.04 | 22.1 | |
| Thyme-RL-7BTool use=Yes2025.11 | 22 | |
| Qwen2.5-VL-7BTool use=No2025.11 | 21.4 | |
| LLaDA-VModel Category=Specialist VLMs2026.04 | 21.2 | |
| MiniCPM-Vversion=4.5, parameters=8B2026.01 | 18.75 | |
| LLaDA-oModel Category=Unified Models2026.04 | 15.7 | |
| Lumina-DiMOOModel Category=Unified Models2026.04 | 13.1 |