{"benchmark_id":"mrcr-1m","benchmark_name":"MRCR 1M","benchmark_description":"MRCR 1M is a variant of the Multi-Round Coreference Resolution benchmark designed for testing extremely long context capabilities with approximately 1 million tokens. It evaluates models' ability to maintain reasoning and attention across ultra-long conversations.","max_score":1.0,"categories":["long_context","reasoning","general"],"modality":"text","total_models":4,"entries":[{"rank":1,"model_id":"deepseek-v4-pro-max","model_name":"DeepSeek-V4-Pro-Max","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.835,"normalized_score":0.835,"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":"2026-04-23","announcement_date":"2026-04-23","multimodal":false,"param_count":1600000000000,"is_new":false},{"rank":2,"model_id":"deepseek-v4-flash-max","model_name":"DeepSeek-V4-Flash-Max","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.787,"normalized_score":0.787,"verified":false,"self_reported":true,"provider_id":"deepseek","input_cost_per_million":0.14,"output_cost_per_million":0.28,"speed_rps":null,"context_window":1048576,"release_date":"2026-04-23","announcement_date":"2026-04-23","multimodal":false,"param_count":284000000000,"is_new":false},{"rank":3,"model_id":"deepseek-v4-flash-0423","model_name":"DeepSeek-V4-Flash-0423","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.769,"normalized_score":0.769,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.1,"output_cost_per_million":0.2,"speed_rps":null,"context_window":1048576,"release_date":"2026-04-23","announcement_date":"2026-04-23","multimodal":false,"param_count":284000000000,"is_new":false},{"rank":4,"model_id":"gemini-2.0-flash-lite","model_name":"Gemini 2.0 Flash-Lite","organization_name":"Google","organization_id":"google","benchmark_score":0.58,"normalized_score":0.58,"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":"2025-02-05","announcement_date":"2025-02-05","multimodal":true,"param_count":null,"is_new":false}]}