all the models — AI benchmark observatory
← Benchmarks

RefCOCO-avg

RefCOCO-avg measures object grounding accuracy averaged across RefCOCO, RefCOCO+, and RefCOCOg benchmarks.

id refcoco-avg · max 100 · 9 models reported

#ModelScore
1Qwen3.6 Plus
Alibaba Cloud / Qwen Team
0.94
2Qwen3.6-27B
Alibaba Cloud / Qwen Team · open
0.93
3Qwen3 VL 235B A22B Thinking
Alibaba Cloud / Qwen Team · open
0.92
4Qwen3.6-35B-A3B
Alibaba Cloud / Qwen Team · open
0.92
5Qwen3.5-122B-A10B
Alibaba Cloud / Qwen Team · open
0.91
6Qwen3.5-27B
Alibaba Cloud / Qwen Team · open
0.91
7Qwen3.5-35B-A3B
Alibaba Cloud / Qwen Team · open
0.89
8LFM2.5-VL-3B
Liquid AI · open
0.88