OpenVLM MMBench V1.1 CN - Physical Relation: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 284 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL3-78B | 80 |
| 2 | InternVL3-38B | 76 |
| 3 | GPT-5 | 72 |
| 4 | InternVL2.5-78B | 72 |
| 5 | InternVL3-8B | 70.7 |
| 6 | GPT-5 Mini (2025-08-07) | 70.7 |
| 7 | Gemini 2.5 Pro | 65.3 |
| 8 | Qwen 2 VL 72B | 62.7 |
| 9 | GPT-4.1 Mini | 58.7 |
| 10 | GPT-5 Nano | 58.7 |
| 11 | Qwen 2 VL 7B | 57.3 |
| 12 | Gemini 1.5 Pro (002) | 57.3 |
| 13 | GPT-4.1 (2025-04-14) | 57.3 |
| 14 | GPT-4.5 | 57.3 |
| 15 | InternVL2-8B | 56 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmbench-v1-1-cn-physical-relation · How It Works · Data refreshed daily, snapshot 2026-09-05.