llava-onevision-qwen2-72B-ov-hf: benchmark results

Provider: LLaVA Team.

Aggregate Rating 1512 ± 37, rank #2118 of 10552 rated models, from 17 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
CoreCognition Image + Video Boundary81.84Accuracy (%)87.8
CoreCognition Image + Video Mechanicalreasoning66.31Accuracy (%)80.7
CoreCognition Image + Video Intentionality78.04Accuracy (%)77.3
CoreCognition Image + Video Hierarchy63.27Accuracy (%)73.9
CoreCognition Image + Video Permanence40.36Accuracy (%)55
CoreCognition Image + Video Mean54.46Accuracy (%)53.8
CoreCognition Image + Video Conservation48.35Accuracy (%)50.4
CoreCognition Image + Video Toolusing54.69Accuracy (%)50.4
CoreCognition Image + Video Continuity57.85Accuracy (%)49.6
CoreCognition Image + Video Perceptualconstancy64.63Accuracy (%)47.1
CapArena-Auto-12.3CapArena-Auto Score (avg)42.9
CapArena-Auto vs CogVLM-19B-6CapArena-Auto Score (vs CogVLM-19B)42.9

Interactive version: theaggregate.ai/model?slug=llava-onevision-qwen2-72b-ov-hf · How It Works · Data refreshed daily, snapshot 2026-11-16.