List your company

GUI planning semantic consistency on GUIGuard-Bench Android 1.0

3.45Android Avg.Gemini-3-Pro
1.18281.77142.362.9486Jan 26, 2026

Evaluation results

MethodLinks
Gemini-3-ProCategory=Closed-source VLM2026.013.45
Claude-Sonnet-4.5Category=Closed-source VLM2026.013.44
GPT-5.2Category=Closed-source VLM2026.013.33
Qwen3-VL-235B-A22BCategory=Open-source VLM2026.012.92
GUI-Owl-7BCategory=GUI Agent2026.012.63
Fara-7BCategory=GUI Agent2026.012.34
UI-Tars-1.5-7BCategory=GUI Agent2026.011.27