Share your thoughts, 1 month free Claude Pro on usSee more

Visual Perception on V*

89Score

Step-GUI-8B

Updated 5d ago

Evaluation Results

Method	Links
Step-GUI-8B 2025.12		89
VL-Scaler-MiMO 2026.05		89
DeepEyes 2026.05		88.9
VL-Scaler 2026.05		86.6
Qwen3-VL 8B-Instruct 2025.12		86.4
Qwen3-VL-30B A3B-Instruct 2025.12		86.4
Qwen2.5-VL-72B 2025.12		85.9
Pixel Reasoner 2026.05		84.3
Qwen3-VL-8B-Thinking + DiG 2025.12		81.2
Qwen2.5-VL-Instruct 2026.05		81.2
Qwen2.5-VL-7B 2026.04		80.1
Step-GUI-4B 2025.12		79.1
Qwen3-VL-4B-Thinking 2025.12		79.1
Qwen3-VL-4B-Thinking + DiG 2025.12		79.1
Qwen3-VL-8B-Thinking 2025.12		79.1
MiMO-VL-Instruct 2026.05		78.6
Qwen2.5-VL-7B 2025.12		77
InternVL2.5-78B 2025.12		75.9
Llava-OV 2026.05		72.8
LoMo 2026.05		71.86
Qwen2.5-VL-Instruct 2026.05		71.4
VL-Rethinker 2026.05		68.2
Standard SFT 2026.05		68.13
BAGEL 2026.04		67.5
GPT5-nano minimal 2025.12		66.5
OpenVLThinker 2026.05		65.2
LoMo 2026.05		64.73
mPLUG-Owl3 2026.05		64.5
LLaDA2.0-Uni 2026.04		61.8
Standard SFT 2026.05		61.58
R1-VL 2026.05		60.3
LLaDA-o 2026.04		57.6
Lumina-DiMOO 2026.04		52.7
LoMo 2026.05		51.7
InternVL-U 2026.04		51.3
GPT-4o-mini 2026.05		50.8
Standard SFT 2026.05		47.71
GPT-4o 2026.05		45
LLaDA-V 2026.04		41.8
Docopilot 2026.05		40.1
LoMo 2026.05		36.13
Standard SFT 2026.05		8.12