List your company

Visual Reasoning on VisualPuzzles OOD (test)

47.95Overall AccuracyGemini-3.1Pro
30.63435.129539.62544.1205May 12, 2026

Evaluation results

MethodLinks
Gemini-3.1Pro2026.0547.9560.6961.145228.2338.11
GPT-4o2026.0541.349.258.34927.326.2
Monet2026.0535.0245.830.8147.526.7925.52
SCOLAR-7BParameters=7B2026.0534.4233.5929.864929.6731.82
Deepeyes2026.0532.9637.7927.014126.7927.01
SCOLAR(SFT only)Mode=SFT only2026.0532.8837.7929.3844.528.7125.87
Qwen2.5-VLParameters=7B2026.0532.7137.0221.847.526.3221.8
PangeaParameters=7B2026.0531.332.423.738.528.732.5