MMMU ↗80.6
Qwen3-VL-235B-A22B · MMMU · Qwen report ↗
Qwen / MODEL EVIDENCE
Qwen3-VL-235B-A22B
11 benchmarks · 12 observations · 1 sources
Scores at a glance
11 benchmarks · large dot: Qwen3-VL-235B-A22B; small dots: closest models where reported
MMMU-Pro ↗69.3
Qwen3-VL-235B-A22B · MMMU-Pro · Qwen report ↗
MathVision ↗74.6
Qwen3-VL-235B-A22B · MathVision · Qwen report ↗
MathVista ↗85.8
Qwen3-VL-235B-A22B · Mathvista(mini) · Qwen report ↗
Qwen3-VL-235B-A22B · ZEROBench · Qwen report ↗
CharXiv ↗66.1
Qwen3-VL-235B-A22B · CharXiv(RQ) · Qwen report ↗
OCRBench ↗87.5
Qwen3-VL-235B-A22B · OCRBench · Qwen report ↗
Video-MME ↗83.8
Qwen3-VL-235B-A22B · VideoMME(w sub.) · Qwen report ↗
1 more reported result
79.0 · Qwen3-VL-235B-A22B · VideoMME(w/o sub.) · Qwen report · Qwen3.5 technical model card · vision ↗
LVBench ↗63.6
Qwen3-VL-235B-A22B · LVBench · Qwen report ↗
ScreenSpot-Pro ↗62.0
Qwen3-VL-235B-A22B · ScreenSpot Pro · Qwen report ↗
Qwen3-VL-235B-A22B · OSWorld-Verified · Qwen report ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
Claude 4.5 Opus →Anthropic
11 of 11 shared benchmark comparisons
GPT-5.2 →OpenAI
10 of 11 shared benchmark comparisons
Kimi K2.5 →Kimi
10 of 11 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Reports featuring this model
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
Names used in source tables
Qwen3-VL-235B-A22B