{"benchmark_id":"recreationbench","benchmark_name":"RecreationBench","benchmark_description":"RecreationBench is Qwen's long-horizon application-recreation benchmark for hybrid agents across Ubuntu, macOS, Windows, Android, and the web.","max_score":1.0,"categories":["multimodal","agents","code","vision"],"modality":"multimodal","total_models":3,"entries":[{"rank":1,"model_id":"qwen3.8-flash","model_name":"Qwen3.8 Flash","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.499,"normalized_score":0.499,"verified":false,"self_reported":true,"provider_id":"novita","input_cost_per_million":0.15,"output_cost_per_million":0.47,"speed_rps":null,"context_window":1000000,"release_date":"2026-08-26","announcement_date":"2026-08-26","multimodal":true,"param_count":125000000000,"is_new":false},{"rank":2,"model_id":"qwen3.8-flash-next","model_name":"Qwen3.8-Flash-Next","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.499,"normalized_score":0.499,"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":"2026-08-26","announcement_date":"2026-08-26","multimodal":true,"param_count":125000000000,"is_new":false},{"rank":3,"model_id":"qwen3.8-27b","model_name":"Qwen3.8-27B","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.471,"normalized_score":0.471,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.4,"output_cost_per_million":3.0,"speed_rps":null,"context_window":262144,"release_date":"2026-08-14","announcement_date":"2026-08-14","multimodal":true,"param_count":27781427952,"is_new":false}]}