Open LMM Reasoning - MathVision - statistics — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 100 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Doubao-1.5-Pro | 63.8 |
| 2 | GPT-4.1 Mini | 56.9 |
| 3 | Gemini 1.5 Pro (002) | 56.9 |
| 4 | GPT-4.1 (2025-04-14) | 56.9 |
| 5 | Claude 3.5 Sonnet (20241022) | 55.2 |
| 6 | GPT-4o ChatGPT | 51.7 |
| 7 | InternVL3-78B | 50 |
| 8 | InternVL3-14B | 48.3 |
| 9 | Gemini 2.0 Flash | 48.3 |
| 10 | QVQ-72B-Preview | 46.6 |
| 11 | InternVL2.5-78B | 44.8 |
| 12 | InternVL3-38B | 43.1 |
| 13 | Claude 3.7 Sonnet | 43.1 |
| 14 | Gemma 3 12B | 43.1 |
| 15 | InternVL3-8B | 39.7 |
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-mathvision-statistics · How the rankings work · Data refreshed daily, snapshot 2026-07-22.