List your company

General and Structural Reasoning on Tree-Bench

56.54AccuracyGemini-3-Flash
36.509641.709846.9152.1102Jun 30, 2026

Evaluation results

MethodLinks
Gemini-3-FlashSize=-2026.0656.54
PixelEyesSize=8B2026.0648.4
Qwen-3-VLSize=8B2026.0646.91
PixelEyesSize=4B2026.0645.93
Qwen-3-VLSize=4B2026.0642.71
Qwen-2.5-VLSize=7B2026.0641.48
Pixel-ReasonerSize=7B2026.0640.98
Mini-o3Size=7B2026.0640.25
ThymeSize=7B2026.0639.75
DeepEyesSize=7B2026.0637.28