OpenVLM MathVista - MWP: leaderboard
OpenCompass OpenVLM MathVista math word problem subcategory score.
Metric: Accuracy (%). Source: huggingface.co. Source: huggingface.co. Saturation forecast: Around January 2027. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-78B | 91.4 | #352 |
| 2 | InternVL3-14B | 89.2 | #505 |
| 3 | InternVL3-38B | 88.2 | #405 |
| 4 | Grok 2 (1212) | 85.5 | #496 |
| 5 | Doubao-1.5-Pro | 84.9 | #287 |
| 6 | InternVL3-8B | 83.3 | #620 |
| 7 | Gemini 2.5 Pro | 80.6 | #152 |
| 8 | Taichu-VLR-7B | 80.1 | #526 |
| 9 | GPT-5 | 79 | #98 |
| 10 | Claude 3.7 Sonnet | 78.5 | #249 |
| 11 | GPT-5 Mini (2025-08-07) | 78.5 | #172 |
| 12 | InternVL2.5-78B | 78 | #431 |
| 13 | InternVL2.5-26B-CoT | 78 | #642 |
| 14 | InternVL2.5-78B-CoT | 76.3 | #469 |
| 15 | URSA-8B | 75.3 | #811 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mathvista-mwp · How It Works · Data refreshed daily, snapshot 2026-10-20.