Emu2_chat Chat: benchmark results

Provider: BAAI. Released 2023-12-22.

Aggregate Rating 1350 ± 45, rank #7714 of 10552 rated models, from 16 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
CoreCognition Image + Video Permanence43.79Accuracy (%)68.1
CoreCognition Image + Video Perspective33.24Accuracy (%)58
CoreCognition Image + Video Spatiality39.58Accuracy (%)42
CoreCognition Image + Video Continuity54.94Accuracy (%)37
CoreCognition Image + Video Hierarchy52.99Accuracy (%)34.5
CMMMU24.5Test Overall (%)25
CoreCognition Image + Video Perceptualconstancy57.27Accuracy (%)23.1
CoreCognition Image + Video Toolusing43.66Accuracy (%)22.7
CoreCognition Image + Video Intuitivephysics45.76Accuracy (%)20.2
CoreCognition Image + Video Mechanicalreasoning44.71Accuracy (%)18.9
CoreCognition Image + Video Boundary64.61Accuracy (%)16.8
CoreCognition Image + Video Mean45.67Accuracy (%)16

Interactive version: theaggregate.ai/model?slug=emu2-chat-chat · How It Works · Data refreshed daily, snapshot 2026-11-16.