OpenVLM MMStar - Science and Technology: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 284 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 71.6 |
| 2 | InternVL3-78B | 68.8 |
| 3 | InternVL3-38B | 66.4 |
| 4 | Gemini 2.5 Pro | 65.2 |
| 5 | InternVL2.5-78B | 64 |
| 6 | GPT-5 Mini (2025-08-07) | 62 |
| 7 | InternVL3-8B | 61.6 |
| 8 | GPT-4.5 | 61.6 |
| 9 | GPT-4.1 (2025-04-14) | 60 |
| 10 | GPT-4o ChatGPT | 59.6 |
| 11 | Qwen 2 VL 72B | 58.4 |
| 12 | Gemini 2.0 Flash | 57.6 |
| 13 | Claude 3.7 Sonnet | 52.8 |
| 14 | Claude 3.5 Sonnet (20241022) | 52.4 |
| 15 | Gemini 1.5 Flash (002) | 52.4 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmstar-science-and-technology · How It Works · Data refreshed daily, snapshot 2026-09-05.