{"benchmark_id":"aitz-em","benchmark_name":"AITZ_EM","benchmark_description":"Android-In-The-Zoo (AitZ) benchmark for evaluating autonomous GUI agents on smartphones. Contains 18,643 screen-action pairs with chain-of-action-thought annotations spanning over 70 Android apps. Designed to connect perception (screen layouts and UI elements) with cognition (action decision-making) for natural language-triggered smartphone task completion.","max_score":1.0,"categories":["multimodal","reasoning","agents"],"modality":"multimodal","total_models":3,"entries":[{"rank":1,"model_id":"qwen2.5-vl-72b","model_name":"Qwen2.5 VL 72B Instruct","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.832,"normalized_score":0.832,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2025-01-26","announcement_date":"2025-01-26","multimodal":true,"param_count":72000000000,"is_new":false},{"rank":2,"model_id":"qwen2.5-vl-32b","model_name":"Qwen2.5 VL 32B Instruct","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.831,"normalized_score":0.831,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2025-02-28","announcement_date":"2025-02-28","multimodal":true,"param_count":33500000000,"is_new":false},{"rank":3,"model_id":"qwen2.5-vl-7b","model_name":"Qwen2.5 VL 7B Instruct","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.819,"normalized_score":0.819,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2025-01-26","announcement_date":"2025-01-26","multimodal":true,"param_count":8290000000,"is_new":false}]}