{"benchmark_id":"mrcr-v2","benchmark_name":"MRCR v2","benchmark_description":"MRCR v2 (Multi-Round Coreference Resolution version 2) is an enhanced version of the synthetic long-context reasoning task. It extends the original MRCR framework with improved evaluation criteria and additional complexity for testing models' ability to maintain attention and reasoning across extended contexts.","max_score":1.0,"categories":["long_context","reasoning","general"],"modality":"text","total_models":3,"entries":[{"rank":1,"model_id":"qwen3.7-plus","model_name":"Qwen3.7-Plus","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.917,"normalized_score":0.917,"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":"2026-05-31","announcement_date":"2026-05-31","multimodal":true,"param_count":null,"is_new":false},{"rank":2,"model_id":"diffusiongemma-26b-a4b-it","model_name":"DiffusionGemma 26B-A4B","organization_name":"Google","organization_id":"google","benchmark_score":0.32,"normalized_score":0.32,"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":"2026-06-10","announcement_date":"2026-06-10","multimodal":true,"param_count":25200000000,"is_new":false},{"rank":3,"model_id":"gemini-2.5-flash-lite","model_name":"Gemini 2.5 Flash-Lite","organization_name":"Google","organization_id":"google","benchmark_score":0.166,"normalized_score":0.166,"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":"2025-06-17","announcement_date":"2025-06-17","multimodal":true,"param_count":null,"is_new":false}]}