List your company
ResearchFormerly WizWand

GUI Grounding on UI-Vision (Accuracy)

58.7AccuracyGROUNDNEXT-7B (SFT)
-1.41214.19429.845.406Nov 10, 2025Dec 18, 2025Jan 26, 2026Mar 5, 2026Apr 13, 2026May 21, 2026Jun 29, 2026
Updated 4mo ago

Evaluation results

MethodLinks
GROUNDNEXT-7B (SFT)Model Scale=7B, Training Paradigm=SFT2025.1158.7
GROUNDNEXT-3B (SFT)Model Scale=3B, Training Paradigm=SFT2025.1158.2
ZoomOnceSize=4B2026.0640.2
MAI-UISize=4B2026.0637
GUI-SDTime/epoch=4.22026.0533.3
OpenCUASize=7B2026.0629.7
Propose-then-Critic-8B2026.0528.9
Qwen3-VL-Instruct + GRPO-GaussianTime/epoch=16.82026.0528.2
Qwen3-VL-Instruct + GRPO-BinaryTime/epoch=16.92026.0527.6
Qwen3-VL-Instruct + GRPO-DistanceTime/epoch=16.72026.0527.5
ZoomUI-7B2026.0527.1
Qwen3-VLSize=32B2026.0626.9
UI-VenusSize=7B2026.0626.5
InfiGUI-G1Size=7B2026.0626.1
HyperClick-7B2026.0525.7
UI-TARS-DPOSize=72B2026.0625.5
JEDI-7B2026.0525.2
Qwen3-VL-Instruct2026.0525.2
JEDI-7BModel Scale=7B, Training Paradigm=SFT2025.1124.8
Qwen3-VLSize=8B2026.0623.3
UI-TARS-1.5Size=7B2026.0622.3
InfiGUI-G1Size=3B2026.0622
GUI-Actor-7BModel Scale=7B, Training Paradigm=SFT2025.1121.9
PhiGround-4B-7CModel Scale=4B, Training Paradigm=SFT2025.1120.5
GUI-Actor-3BModel Scale=3B, Training Paradigm=SFT2025.1119.7
JEDI-3BModel Scale=3B, Training Paradigm=SFT2025.1118.7
TongUI-7B2026.0518
UI-TARS-7B2026.0517.6
UI-TARSSize=7B2026.0617.6
Qwen2.5-VL-7B (Agent mode)Model Scale=7B, Evaluation Mode=Agent mode, Training Paradigm=SFT2025.1116.5
Aguvis-7BModel Scale=7B, Training Paradigm=SFT2025.1113.7
UGround-V1-7BModel Scale=7B, Training Paradigm=SFT2025.1112.9
UGroundSize=7B2026.0612.9
OS-Atlas-7BModel Scale=7B, Training Paradigm=SFT2025.119
OS-AtlasSize=7B2026.069
Qwen2.5-VL-3BModel Scale=3B, Training Paradigm=SFT2025.116.3
Qwen2.5-VL-3B (Agent mode)Model Scale=3B, Evaluation Mode=Agent mode, Training Paradigm=SFT2025.116.3
Qwen2.5-VL-7BModel Scale=7B, Training Paradigm=SFT2025.110.9