Qwen2-VL: benchmark results

Provider: Alibaba.

Aggregate Rating 1529 ± 46, rank #1959 of 10916 rated models, from 19 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Video-MME - Long Videos (w/ subs)74.3Accuracy (%)92
Video-MME - Medium Videos (w/ subs)76.8Accuracy (%)92
Video-MME - Medium Videos (w/o subs)71.3Accuracy (%)90
Video-MME - Long Videos (w/o subs)62.2Accuracy (%)84
VideoMME w sub.77.8Score (imported)83.6
Video-MME - Short Videos (w/o subs)80.1Accuracy (%)83
Video-MME - Short Videos (w/ subs)82.2Accuracy (%)82
V-STaR - Long mAM14.09Score69.2
V-STaR - Long mLGM17.53Score69.2
V-STaR - Medium mAM18.47Score61.5
V-STaR - Medium mLGM21.22Score38.5
VisualWebArena2.9Success Rate (%)25.8

Interactive version: theaggregate.ai/model?slug=qwen2-vl · How It Works · Data refreshed daily, snapshot 2026-11-08.