OpenVLM MMMU - Design — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 284 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 Mini | 90 |
| 2 | Gemini 1.5 Pro (002) | 90 |
| 3 | GPT-5 (2025-08-07) | 90 |
| 4 | Ovis2-8B | 86.7 |
| 5 | GPT-4.5 | 86.7 |
| 6 | Gemini 2.5 Pro | 83.3 |
| 7 | InternVL3-78B | 83.3 |
| 8 | InternVL3-38B | 83.3 |
| 9 | InternVL3-14B | 83.3 |
| 10 | Claude 3.7 Sonnet | 83.3 |
| 11 | Gemini 1.5 Pro | 83.3 |
| 12 | GPT-4.1 (2025-04-14) | 83.3 |
| 13 | Llama 3.2 90B Vision Instruct | 83.3 |
| 14 | GPT-4o ChatGPT | 83.3 |
| 15 | GPT-5 Mini (2025-08-07) | 83.3 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmmu-design · How the rankings work · Data refreshed daily, snapshot 2026-07-22.