{"benchmark_id":"piqa","benchmark_name":"PIQA","benchmark_description":"PIQA (Physical Interaction: Question Answering) is a benchmark dataset for physical commonsense reasoning in natural language. It tests AI systems' ability to answer questions requiring physical world knowledge through multiple choice questions with everyday situations, focusing on atypical solutions inspired by instructables.com. The dataset contains 21,000 multiple choice questions where models must choose the most appropriate solution for physical interactions.","max_score":1.0,"categories":["physics","reasoning","general"],"modality":"text","total_models":11,"entries":[{"rank":1,"model_id":"phi-3.5-moe-instruct","model_name":"Phi-3.5-MoE-instruct","organization_name":"Microsoft","organization_id":"microsoft","benchmark_score":0.886,"normalized_score":0.886,"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":"2024-08-23","announcement_date":"2024-08-23","multimodal":false,"param_count":60000000000,"is_new":false},{"rank":2,"model_id":"hermes-3-70b","model_name":"Hermes 3 70B","organization_name":"Nous Research","organization_id":"nous-research","benchmark_score":0.8444,"normalized_score":0.8444,"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":"2024-08-15","announcement_date":"2024-08-15","multimodal":false,"param_count":70000000000,"is_new":false},{"rank":3,"model_id":"gemma-2-27b-it","model_name":"Gemma 2 27B","organization_name":"Google","organization_id":"google","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":"2024-06-27","announcement_date":"2024-06-27","multimodal":false,"param_count":27200000000,"is_new":false},{"rank":4,"model_id":"gemma-2-9b-it","model_name":"Gemma 2 9B","organization_name":"Google","organization_id":"google","benchmark_score":0.817,"normalized_score":0.817,"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":"2024-06-27","announcement_date":"2024-06-27","multimodal":false,"param_count":9240000000,"is_new":false},{"rank":5,"model_id":"gemma-3n-e4b","model_name":"Gemma 3n E4B","organization_name":"Google","organization_id":"google","benchmark_score":0.81,"normalized_score":0.81,"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":"2025-06-26","announcement_date":"2025-06-26","multimodal":true,"param_count":8000000000,"is_new":false},{"rank":6,"model_id":"gemma-3n-e4b-it-litert-preview","model_name":"Gemma 3n E4B Instructed LiteRT Preview","organization_name":"Google","organization_id":"google","benchmark_score":0.81,"normalized_score":0.81,"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":"2025-05-20","announcement_date":"2025-05-20","multimodal":true,"param_count":1910000000,"is_new":false},{"rank":7,"model_id":"phi-3.5-mini-instruct","model_name":"Phi-3.5-mini-instruct","organization_name":"Microsoft","organization_id":"microsoft","benchmark_score":0.81,"normalized_score":0.81,"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":"2024-08-23","announcement_date":"2024-08-23","multimodal":false,"param_count":3800000000,"is_new":false},{"rank":8,"model_id":"gemma-3n-e2b","model_name":"Gemma 3n E2B","organization_name":"Google","organization_id":"google","benchmark_score":0.789,"normalized_score":0.789,"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":"2025-06-26","announcement_date":"2025-06-26","multimodal":true,"param_count":8000000000,"is_new":false},{"rank":9,"model_id":"gemma-3n-e2b-it-litert-preview","model_name":"Gemma 3n E2B Instructed LiteRT (Preview)","organization_name":"Google","organization_id":"google","benchmark_score":0.789,"normalized_score":0.789,"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":"2025-05-20","announcement_date":"2025-05-20","multimodal":true,"param_count":1910000000,"is_new":false},{"rank":10,"model_id":"phi-4-mini","model_name":"Phi 4 Mini","organization_name":"Microsoft","organization_id":"microsoft","benchmark_score":0.776,"normalized_score":0.776,"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":"2025-02-01","announcement_date":"2025-02-01","multimodal":false,"param_count":3840000000,"is_new":false},{"rank":11,"model_id":"ernie-4.5","model_name":"ERNIE 4.5","organization_name":"Baidu","organization_id":"baidu","benchmark_score":0.552,"normalized_score":0.552,"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":"2025-06-25","announcement_date":"2025-06-25","multimodal":false,"param_count":21000000000,"is_new":false}]}