OpenVLM Video - Video-MME (w/o subs) - Object Reasoning: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 37 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 71.1 |
| 2 | InternVL2.5-78B | 70.9 |
| 3 | GPT-4o (2024-08-06) | 68.9 |
| 4 | Gemini 2.0 Flash | 68.1 |
| 5 | Qwen 2 VL 72B | 67.8 |
| 6 | InternVL3-8B | 63.2 |
| 7 | Gemini 1.5 Flash | 59.7 |
| 8 | Qwen 2 VL 7B | 57.9 |
| 9 | InternVL2-8B | 52.2 |
| 10 | Claude 3.5 Sonnet | 47.4 |
| 11 | Phi-4 Multimodal Instruct | 42.1 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-video-video-mme-w-o-subs-object-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.