OpenVLM VCR — leaderboard
OpenCompass OpenVLM evaluation of VCR: visual clue reasoning over English and Chinese examples, scored with exact match and Jaccard metrics for easy and hard visual reasoning questions.
Metric: Overall Jaccard (%). Source: huggingface.co. 48 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 VL 7B | 75.6 |
| 2 | Qwen 2 VL 2B | 61.3 |
| 3 | Gemini 1.5 Pro | 38.1 |
| 4 | Claude 3.5 Sonnet | 35.6 |
| 5 | MiniCPM-V-2.6 | 33.7 |
| 6 | InternVL2-8B | 28.6 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-vcr · How the rankings work · Data refreshed daily, snapshot 2026-07-22.