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

GUI agent success on WebVoyager (test)

90.6Success Rate

GPT-5 (SoM)

Updated 19d ago

Evaluation Results

Method	Links
GPT-5 (SoM) 2026.05		90.6
Gemini computer-use-preview 2026.05		88.6
Gemini-3-flash (Axtree) 2026.05		85.6
o3 (SoM) 2026.05		79.3
MolmoWeb-8B 2026.05		78.2
MolmoWeb-4B 2026.05		75.2
Gemini-3-flash (Axtree) 2026.05		74.4
Orchard-GUI-4B (SFT + RL) 2026.05		74.1
Fara-7B 2026.05		73.5
OpenAI computer-use-preview 2026.05		70.9
GPT-5 (Axtree) 2026.05		70.6
GLM-4.1V-9B-Thinking 2026.05		66.8
UI-TARS-1.5-7B 2026.05		66.4
GPT-4o (SoM) 2026.05		65.1
Qwen3-VL-235B-A22B-Thinking 2026.05		63.1
Orchard-GUI-4B-SFT 2026.05		60.2
Holo1-7B 2026.05		55.4
Qwen3-VL-4B-Thinking 2026.05		49