{"benchmark_id":"phybench","benchmark_name":"PHYBench","benchmark_description":"PHYBench is a benchmark of real-world physics problems spanning mechanics, electromagnetism, thermodynamics, optics, and modern physics, designed to evaluate physical perception and multi-step quantitative reasoning in large language models.","max_score":1.0,"categories":["physics","reasoning","science"],"modality":"text","total_models":3,"entries":[{"rank":1,"model_id":"hy3","model_name":"Hy3","organization_name":"Tencent","organization_id":"tencent","benchmark_score":0.774,"normalized_score":0.774,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.14,"output_cost_per_million":0.58,"speed_rps":null,"context_window":262144,"release_date":"2026-07-06","announcement_date":"2026-07-06","multimodal":false,"param_count":295000000000,"is_new":false},{"rank":2,"model_id":"seed-2.0-mini","model_name":"Seed 2.0 Mini","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.56,"normalized_score":0.56,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.1,"output_cost_per_million":0.4,"speed_rps":null,"context_window":256000,"release_date":"2026-02-17","announcement_date":"2026-02-17","multimodal":true,"param_count":null,"is_new":false},{"rank":3,"model_id":"seed-1.8","model_name":"Seed 1.8","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.41,"normalized_score":0.41,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.25,"output_cost_per_million":2.0,"speed_rps":null,"context_window":256000,"release_date":"2026-02-17","announcement_date":"2026-02-17","multimodal":true,"param_count":null,"is_new":false}]}