OpenVLM MMT-Bench - VC: leaderboard
Metric: Score (%). Source: huggingface.co. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 VL 72B | 93 |
| 2 | InternVL3-38B | 93 |
| 3 | GPT-4.1 (2025-04-14) | 93 |
| 4 | InternVL3-78B | 92.4 |
| 5 | Llama 3.2 90B Vision Instruct | 89.1 |
| 6 | Grok 2 (1212) | 88.1 |
| 7 | InternVL3-8B | 87.2 |
| 8 | InternVL2-8B | 85.5 |
| 9 | Pixtral-12B | 84.8 |
| 10 | Qwen 2 VL 7B | 84.1 |
| 11 | Gemini 1.5 Flash | 83.4 |
| 12 | Gemini 1.5 Pro | 81.6 |
| 13 | Gemma 3 27B | 79.1 |
| 14 | Llama 3.2 11B Instruct | 78.9 |
| 15 | Gemma 3 12B | 78.5 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmt-bench-vc · How It Works · Data refreshed daily, snapshot 2026-09-05.