OpenVLM Video - MMBench-Video - Single-Instance Fine-grained Perception: leaderboard
Metric: GPT-4 Rating (0-3). Source: huggingface.co. 45 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o (2024-08-06) | 2.24 |
| 2 | Gemini 2.0 Flash | 2.11 |
| 3 | InternVL2.5-78B | 2.1 |
| 4 | Gemini 1.5 Pro | 2.04 |
| 5 | InternVL3-78B | 1.97 |
| 6 | InternVL3-8B | 1.85 |
| 7 | Qwen 2 VL 72B | 1.84 |
| 8 | Gemini 1.5 Flash | 1.69 |
| 9 | Qwen 2 VL 7B | 1.53 |
| 10 | Claude 3.5 Sonnet | 1.39 |
| 11 | InternVL2-8B | 1.37 |
| 12 | Phi-4 Multimodal Instruct | 0.86 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-video-mmbench-video-single-instance-fine-grained-perception · How It Works · Data refreshed daily, snapshot 2026-09-19.