OpenFlamingo v2: benchmark results

Provider: Allen AI. Released 2023-05-17. Access: Open.

Aggregate Rating 1164 ± 51, rank #9618 of 10489 rated models, from 402 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
OpenVLM MMT-Bench - Sign Language Recognition50Score (%)96.8
OpenVLM MMT-Bench - Behavior Anomaly Detection65Score (%)91.5
OpenVLM MMT-Bench - Action Quality Assessment30Score (%)84.7
OpenVLM COCO Captions14.9CIDEr82.4
OpenVLM MMT-Bench - Color Contrast65Score (%)81.8
OpenVLM MMT-Bench - Small Object Detection50Score (%)77.7
OpenVLM MMT-Bench - Screenshot-to-Code45Score (%)76.7
OpenVLM MMT-Bench - Raven Progressive Matrices20Score (%)76.2
OpenVLM MMT-Bench - Color Assimilation65Score (%)75
OpenVLM MMT-Bench - Image Matting25Score (%)72.1
OpenVLM MMT-Bench - Spot The Diff50Score (%)71.8
OpenVLM MMT-Bench - Jigsaw Puzzle Solving30Score (%)63.3

Interactive version: theaggregate.ai/model?slug=openflamingo-v2 · How It Works · Data refreshed daily, snapshot 2026-11-20.