List your company

GUI Element Localization on ScreenSpot Pro

34.5AccuracyKimi-VL-16B-MoE
-0.5488.55117.6526.749May 29, 2025

Evaluation results

MethodLinks
Kimi-VL-16B-MoEBackbone=16B-MoE2025.0534.5
ViGoRL-7bBackbone=7b2025.0533.1
ViGoRL-3bBackbone=3b, Feedback Loop=Multi-turn RL2025.0532.3
UGround-V1-7BBackbone=7B2025.0531.1
ViGoRL-3bBackbone=3b2025.0531.1
Qwen2.5-VL-3B + Vanilla GRPOBackbone=3B, Training Protocol=Vanilla GRPO2025.0529
Qwen2.5-VL-7BBackbone=7B2025.0529
UI-TARS-2BBackbone=2B2025.0527.7
Qwen2.5-VL-3B + SFT directBackbone=3B, Training Protocol=SFT direct2025.0525
Qwen2.5-VL-3BBackbone=3B2025.0523.9
OS-Atlas-7BBackbone=7B2025.0518.9
UI-R1-3BBackbone=3B2025.0517.8
Claude Comp. Use2025.0517.1
UGround-7BBackbone=7B2025.0516.5
CogAgent-18BBackbone=18B2025.057.7
ShowUI-2BBackbone=2B2025.057.7
OS-Atlas-4BBackbone=4B2025.053.7
Qwen2-VL-7BBackbone=7B2025.051.6
SeeClick2025.051.1
GPT-4o2025.050.8