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

Step success rate on AndroidControl (Cat-Unseen)

61.2Step Success Rate (SSR)

T3A* (Qwen2.5-7B-Instruct-ASL)

Updated 11d ago

Evaluation Results

Method	Links
T3A* (Qwen2.5-7B-Instruct-ASL) 2025.06		61.2
M3A (GPT-4o) 2025.06		60.8
T3A* (Qwen2.5-7B-Instruct-SFT) 2025.06		59.7
T3A (GPT-4o) 2025.06		56.5
M3A (Gemini-2.5-Flash) 2025.06		47.9
T3A* (Gemini-2.5-Flash) 2025.06		44.7
T3A (Gemini-2.5-Flash) 2025.06		44.4
SeeAct (Gemini-2.5-Flash) 2025.06		33.7
SeeAct (GPT-4o) 2025.06		30.6
T3A* (Qwen2.5-7B-Instruct) 2025.06		26.7