RBDashMM3-DPO-38B: benchmark results

Provider: Other.

Aggregate Rating 1697 ± 46, rank #361 of 10916 rated models, from 9 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MM-Vet81.2Score97
MathVista79.3Accuracy (%)95.5
AI2D88.7Accuracy (%)95.2
OpenVLM OCRBench90.1Score (normalized)94.4
HallusionBench61.6Overall Score94
MMStar72.1Accuracy (%)92.5
OpenCompass Multimodal - Average78.9Average Score91.3
OpenCompass Multimodal - MMBench87.2Score (%)90.4
OpenCompass Multimodal - MMMU70.9Score (%)83.7

Interactive version: theaggregate.ai/model?slug=rbdashmm3-dpo-38b · How It Works · Data refreshed daily, snapshot 2026-11-08.