Open LMM Reasoning - MathVision - descriptive geometry: leaderboard
The "MathVision - descriptive geometry" subtask of Open LMM Reasoning. A task from the Open LMM Reasoning leaderboard, which evaluates large multimodal models on multi-step visual, mathematical and logical reasoning problems.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 100 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o ChatGPT | 30.8 | |
| 2 | InternVL3-38B | 29.8 | #405 |
| 3 | GPT-4.1 (2025-04-14) | 27.9 | #250 |
| 4 | Doubao-1.5-Pro | 27.9 | #287 |
| 5 | GPT-4.1 Mini | 26.9 | #355 |
| 6 | Gemma 3 27B | 26 | #607 |
| 7 | InternVL3-8B | 25 | #620 |
| 8 | Gemini 2.0 Flash | 25 | #342 |
| 9 | Qwen 2 VL 72B | 25 | #460 |
| 10 | VLM-R1-3B-Math-0305 | 25 | #743 |
| 11 | Claude 3.5 Sonnet (20241022) | 24 | #291 |
| 12 | InternVL3-78B | 24 | #352 |
| 13 | Gemini 1.5 Pro (002) | 23.1 | #349 |
| 14 | Qwen 2.5 VL 7B | 23.1 | #643 |
| 15 | Moonshot v1 8K | 23.1 | #617 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-mathvision-descriptive-geometry · How It Works · Data refreshed daily, snapshot 2026-10-20.