Open LMM Reasoning - MathVision - descriptive geometry: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 100 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o ChatGPT | 30.8 |
| 2 | InternVL3-38B | 29.8 |
| 3 | GPT-4.1 (2025-04-14) | 27.9 |
| 4 | Doubao-1.5-Pro | 27.9 |
| 5 | GPT-4.1 Mini | 26.9 |
| 6 | Gemma 3 27B | 26 |
| 7 | InternVL3-8B | 25 |
| 8 | Qwen 2 VL 72B | 25 |
| 9 | Gemini 2.0 Flash | 25 |
| 10 | VLM-R1-3B-Math-0305 | 25 |
| 11 | InternVL3-78B | 24 |
| 12 | Claude 3.5 Sonnet (20241022) | 24 |
| 13 | Gemini 1.5 Pro (002) | 23.1 |
| 14 | VLAA-Thinker-Qwen2.5VL-3B | 23.1 |
| 15 | Qwen 2 VL 7B | 22.1 |
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-mathvision-descriptive-geometry · How It Works · Data refreshed daily, snapshot 2026-09-05.