List your company
ResearchFormerly WizWand

Visual Perception Reasoning on V* Bench

89.01ScoreZoomEyes
47.066857.955968.84579.7341Sep 26, 2025Oct 21, 2025Nov 16, 2025Dec 11, 2025Jan 6, 2026Jan 31, 2026Feb 26, 2026
Updated 7mo ago

Evaluation results

MethodLinks
ZoomEyesModel=Qwen2.5VL-3B, Training-free=true2026.0289.01
AdaFocusModel=Qwen2.5VL-3B, Training-free=true2026.0286.38
Pixel ReasonerModel=Qwen2.5VL-3B, Training-free=false2026.0284.82
ERGOPixel Constraint=1280x28x28, Post-training Category=Efficiency-oriented Post Training Methods2025.0983.8
ZoomEyesModel=LLaVA-v1.5-7B, Training-free=true2026.0283.25
ERGOPixel Constraint=640x28x28, Post-training Category=Efficiency-oriented Post Training Methods2025.0981.7
MiniO3Pixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0981.2
DeepEyesPixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0978.5
MGPOPixel Constraint=1280x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=reproduction with their code using our data2025.0977.5
Qwen2.5-VL-7B-Inst.Pixel Constraint=16384x28x282025.0977
TreeVGRPixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0976.4
BaselineModel=Qwen2.5VL-3B, Training-free=true2026.0275.9
MLLMs-KnowModel=Qwen2.5VL-3B, Training-free=true2026.0275.9
MiniO3Pixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0974.9
PixelReasonerPixel Constraint=1280x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0974.5
VisionThinkPixel Constraint=1280x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0973.8
MGPOPixel Constraint=640x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=reproduction with their code using our data2025.0967.5
PixelReasonerPixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0967.2
TreeVGRPixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0967
Qwen2.5-VL-7B-Inst.Pixel Constraint=1280x28x282025.0964.9
DeepEyesPixel Constraint=640x28x28, Post-training Category=Non-efficiency-oriented Post Training Methods2025.0964.9
AdaFocusModel=LLaVA-v1.5-7B, Training-free=true2026.0262.8
VisCropModel=LLaVA-v1.5-7B, Training-free=true2026.0262.3
VisionThinkPixel Constraint=640x28x28, Post-training Category=Efficiency-oriented Post Training Methods, Inference Pipeline=inference with original pipeline2025.0961.8
DC2Model=LLaVA-v1.5-7B, Training-free=true2026.0257.6
Qwen2.5-VL-7B-Inst.Pixel Constraint=640x28x282025.0956.5
MLLMs-KnowModel=LLaVA-v1.5-7B, Training-free=true2026.0256.02
BaselineModel=LLaVA-v1.5-7B, Training-free=true2026.0248.68