OpenVLM MMT-Bench - Science: leaderboard
Metric: Score (%). Source: huggingface.co. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 75 |
| 2 | InternVL3-38B | 66.7 |
| 3 | Claude 3.7 Sonnet | 58.3 |
| 4 | GPT-4.1 (2025-04-14) | 58.3 |
| 5 | Gemini 1.5 Pro | 50 |
| 6 | Grok 2 (1212) | 50 |
| 7 | Gemma 3 12B | 50 |
| 8 | InternVL2-8B | 41.7 |
| 9 | Qwen 2 VL 72B | 41.7 |
| 10 | Aquila-VL-2B | 41.7 |
| 11 | Gemma 3 27B | 41.7 |
| 12 | Claude 3.5 Sonnet | 41.7 |
| 13 | Gemini 1.5 Flash | 41.7 |
| 14 | Claude 3 Haiku | 41.7 |
| 15 | GPT-4.1 Mini | 33.3 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-science · How It Works · Data refreshed daily, snapshot 2026-09-05.