List your company

Mobile UI Task Automation on AndroidWorld (full)

82.79Success Rate (Easy)Gemini-3-Flash
49.541258.173166.80575.4369May 11, 2026

Evaluation results

MethodLinks
Gemini-3-FlashAgent Framework=M3A, World-model Feedback=Full Text2026.0582.7961.8131.5867.8961,207.1812.171,042.4
Gemini-3-FlashAgent Framework=M3A, World-model Feedback=Delta Text2026.0581.1558.3335.5366.5956,767.7311.69980.39
GPT-5.4Agent Framework=M3A, World-model Feedback=Delta Text2026.0576.2359.7231.5863.7948,55712.461,017.37
GPT-5.4Agent Framework=M3A, World-model Feedback=None (Baseline)2026.0572.6841.6729.8256.032,178.8911.44220.6
Gemini-3-FlashAgent Framework=M3A, World-model Feedback=Code2Image2026.0572.1344.4410.5353.44217,123.2912.722,714.06
GPT-5.4Agent Framework=M3A, World-model Feedback=Full Text2026.0572.1354.1734.2160.3456,754.3412.611,353.21
GPT-5.4Agent Framework=M3A, World-model Feedback=Code2Image2026.0569.6756.9426.3258.62224,245.8612.313,107.67
Qwen3-VL-32BAgent Framework=M3A, World-model Feedback=Delta Text2026.0564.7550.6921.0553.2347,497.2113.87958.14
Qwen3-VL-32BAgent Framework=M3A, World-model Feedback=Full Text2026.0563.115021.0552.1651,497.5314.11,095.01
Qwen3-VL-32BAgent Framework=M3A, World-model Feedback=Code2Image2026.0563.1143.0618.4249.57209,073.9513.892,490.72
Qwen3-VL-32BAgent Framework=M3A, World-model Feedback=None (Baseline)2026.0562.5744.924.5650.861,907.0113.78553.32
Gemini-3-FlashAgent Framework=M3A, World-model Feedback=None (Baseline)2026.0562.2944.4424.5650.571,238.8611.05290.61
Qwen3-VL-8BAgent Framework=M3A, World-model Feedback=Delta Text2026.0559.0134.7217.144.6154,461.7913.02439.93
Qwen3-VL-8BAgent Framework=M3A, World-model Feedback=Code2Image2026.0554.9231.9413.1640.95214,360.6613.442,600.86
Qwen3-VL-8BAgent Framework=M3A, World-model Feedback=None (Baseline)2026.0554.6427.7716.6640.081,698.1313.69144.56
Qwen3-VL-8BAgent Framework=M3A, World-model Feedback=Full Text2026.0550.8233.3311.8439.0158,548.3213.871,469.09