List your company
ResearchFormerly WizWand

Visual Reasoning on VisualProbe

42.7ScoreERGO
9.62818.21426.835.386Sep 26, 2025
Updated 7mo ago

Evaluation results

MethodLinks
ERGOPixel Constraint=1280x28x28, Post-training Category=Efficiency-oriented Post Training Methods2025.0942.7
DeepEyesPixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0939
MiniO3Pixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0939
ERGOPixel Constraint=640x28x28, Post-training Category=Efficiency-oriented Post Training Methods2025.0935
MGPOPixel Constraint=1280x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=reproduction with their code using our data2025.0933.4
PixelReasonerPixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0931.8
MiniO3Pixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0931.7
MGPOPixel Constraint=640x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=reproduction with their code using our data2025.0929.7
Qwen2.5-VL-7B-Inst.Pixel Constraint=16384x28x282025.0929.2
DeepEyesPixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0926.8
TreeVGRPixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0926.6
TreeVGRPixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0924.1
PixelReasonerPixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0923.9
VisionThinkPixel Constraint=640x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0917.9
VisionThinkPixel Constraint=1280x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0917.3
Qwen2.5-VL-7B-Inst.Pixel Constraint=1280x28x282025.0915.2
Qwen2.5-VL-7B-Inst.Pixel Constraint=640x28x282025.0910.9