Open LMM Reasoning - MathVision - graph theory: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 100 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 Mini | 41.1 |
| 2 | GPT-4o ChatGPT | 36.7 |
| 3 | GPT-4.1 (2025-04-14) | 31.1 |
| 4 | InternVL3-78B | 28.9 |
| 5 | Gemini 1.5 Pro (002) | 28.9 |
| 6 | Gemini 2.0 Flash | 28.9 |
| 7 | GPT-4.1 Nano | 26.7 |
| 8 | Gemma 3 27B | 26.7 |
| 9 | Qwen 2 VL 72B | 25.6 |
| 10 | Doubao-1.5-Pro | 25.6 |
| 11 | QVQ-72B-Preview | 25.6 |
| 12 | InternVL3-38B | 24.4 |
| 13 | Claude 3.7 Sonnet | 24.4 |
| 14 | InternVL2.5-78B-CoT | 24.4 |
| 15 | Claude 3.5 Sonnet (20241022) | 23.3 |
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-mathvision-graph-theory · How It Works · Data refreshed daily, snapshot 2026-09-05.