OpenVLM Video - Video-MME (w/o subs) - Action Recognition: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 37 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 72.2 |
| 2 | InternVL2.5-78B | 70.3 |
| 3 | Gemini 2.0 Flash | 70 |
| 4 | GPT-4o (2024-08-06) | 66.8 |
| 5 | Qwen 2 VL 72B | 65.5 |
| 6 | InternVL3-8B | 62.6 |
| 7 | Qwen 2 VL 7B | 60.4 |
| 8 | Gemini 1.5 Flash | 59.4 |
| 9 | InternVL2-8B | 48.9 |
| 10 | Claude 3.5 Sonnet | 48.6 |
| 11 | Phi-4 Multimodal Instruct | 40.9 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-video-video-mme-w-o-subs-action-recognition · How It Works · Data refreshed daily, snapshot 2026-09-19.