{"benchmark_id":"swe-atlas-codebase-qna","benchmark_name":"SWE Atlas - Codebase QnA","benchmark_description":"SWE Atlas - Codebase QnA evaluates a model's ability to answer questions about real codebases, measuring repository-level comprehension and the ability to reason about code structure, behavior, and intent across an entire project.","max_score":1.0,"categories":["agents","code"],"modality":"text","total_models":2,"entries":[{"rank":1,"model_id":"laguna-s-2.1","model_name":"Laguna S 2.1","organization_name":"Poolside","organization_id":"poolside","benchmark_score":0.462,"normalized_score":0.462,"verified":false,"self_reported":true,"provider_id":"poolside","input_cost_per_million":0.1,"output_cost_per_million":0.2,"speed_rps":null,"context_window":1048576,"release_date":"2026-07-21","announcement_date":"2026-07-21","multimodal":false,"param_count":118000000000,"is_new":false},{"rank":2,"model_id":"minimax-m3","model_name":"MiniMax M3","organization_name":"MiniMax","organization_id":"minimax","benchmark_score":0.379,"normalized_score":0.379,"verified":false,"self_reported":true,"provider_id":"novita","input_cost_per_million":0.3,"output_cost_per_million":1.2,"speed_rps":null,"context_window":1000000,"release_date":"2026-06-01","announcement_date":"2026-06-01","multimodal":true,"param_count":428000000000,"is_new":false}]}