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

Mobile Agent Interaction on MobileIAR Chinese users

51.97SR

OS-Atlas-7B-Pro

Updated 13d ago

Evaluation Results

Method	Links
OS-Atlas-7B-Pro 2025.08		51.97	81.64	47.82
UI-TARS-7B-SFT 2025.08		51.04	75.86	48.7
UI-TARS-1.5-7B 2025.08		49.8	77.01	46.75
UI-TARS-7B-DPO 2025.08		46.45	73.73	43.46
UI-TARS-1.5-7B 2025.08		46.22	76.42	39.18
UI-TARS-7B-DPO 2025.08		45.07	70.74	37.19
UI-TARS-7B-SFT 2025.08		44.86	75	36.86
GUI-owl-7B 2025.08		44.22	73.26	38.49
GPT-4o + OCR model 2025.08		44.19	78.06	41.4
Qwen3-VL-8B 2025.08		43.01	70.21	37.36
OS-Atlas-7B-Pro 2025.08		42.22	76.92	35.67
GUI-owl-7B 2025.08		39.32	70.21	35.07
Qwen2.5-VL-7B 2025.08		38.27	47.27	36.13
Qwen3-VL-8B 2025.08		38.08	65.57	32.09
GPT-4o + OCR model 2025.08		37.13	77.42	31.18
Qwen-VL-max + OCR model 2025.08		24.04	84.42	21.34
Qwen-VL-max + OCR model 2025.08		22	81.55	19.04
Qwen2.5-VL-7B 2025.08		19.29	23.52	18.13
GLM-4v + OCR model 2025.08		4.05	73.71	3.03
GLM-4v + OCR model 2025.08		3.11	72.97	2.47