Multimodal mathematical reasoning on MathVerse (vision)
76.39Pass@1 AccuracyMiMo-VL-7B-RL-2508
Evaluation results
| Method | Links | |
|---|---|---|
| MiMo-VL-7B-RL-2508Series=MiMo-VL-7B2026.02 | 76.39 | |
| MiMo-VL-7B-DeepVisionSeries=MiMo-VL-7B, Training Algorithm=GSPO2026.02 | 76.26 | |
| MiMo-VL-7B-OpenMMReasonerSeries=MiMo-VL-7B2026.02 | 74.87 | |
| MiMo-VL-7B-MathBookSeries=MiMo-VL-7B2026.02 | 73.6 | |
| MiMo-VL-7B-MM-EurekaSeries=MiMo-VL-7B2026.02 | 73.35 | |
| Qwen3-VL-8B-ThinkingSeries=Qwen3-VL-8B2026.02 | 72.84 | |
| MiMo-VL-7B-SFT-2508Series=MiMo-VL-7B2026.02 | 72.71 | |
| Qwen3-VL-8B-DeepVisionSeries=Qwen3-VL-8B, Training Algorithm=GSPO2026.02 | 72.46 | |
| GPT-5-Nano-HighModel Category=Closed-source Models2026.02 | 70.3 | |
| Gemini-2.5-Flash-LiteModel Category=Closed-source Models2026.02 | 70.3 | |
| Qwen3-VL-8B-InstructSeries=Qwen3-VL-8B2026.02 | 67.38 |