List your company

Visual Reasoning on VisualProbe Medium

40.6AccuracyDeepconf
14.39221.1962834.804Feb 13, 2026

Evaluation results

MethodLinks
DeepconfBackbone=Qwen3-VL Instruct2026.0240.616.7
RTWIBackbone=Qwen3-VL Instruct2026.0240.618.5
SCBackbone=Qwen3-VL Instruct2026.0240.2—
ASCBackbone=Qwen3-VL Instruct2026.0240.221.8
ESCBackbone=Qwen3-VL Instruct2026.0240.28.8
Self-Cer.Backbone=Qwen3-VL Instruct2026.0239.414.5
CISCBackbone=Qwen3-VL Instruct2026.0238.614.7
BaseBackbone=Qwen3-VL Instruct2026.0235.4—
RTWIBackbone=Qwen3-VL Thinking2026.0234.141.9
DeepEyesBackbone=DeepEyes-7B2026.0229.8—
SCBackbone=Qwen3-VL Thinking2026.0229.6—
ASCBackbone=Qwen3-VL Thinking2026.0229.632.6
ESCBackbone=Qwen3-VL Thinking2026.0229.610.9
DeepconfBackbone=Qwen3-VL Thinking2026.0229.641
CISCBackbone=Qwen3-VL Thinking2026.0227.720.4
Self-Cer.Backbone=Qwen3-VL Thinking2026.0227.321.4
BaseBackbone=Qwen3-VL Thinking2026.0225.8—
GPT-4oBackbone=GPT-4o2026.0215.4—