{"benchmark_id":"mrcr-v2-(8-needle)","benchmark_name":"MRCR v2 (8-needle)","benchmark_description":"MRCR v2 (8-needle) is a variant of the Multi-Round Coreference Resolution benchmark that includes 8 needle items to retrieve from long contexts. This tests models' ability to simultaneously track and reason about multiple pieces of information across extended conversations.","max_score":1.0,"categories":["long_context","reasoning","general"],"modality":"text","total_models":23,"entries":[{"rank":1,"model_id":"gemini-3.7-flash","model_name":"Gemini 3.7 Flash","organization_name":"Google","organization_id":"google","benchmark_score":0.97,"normalized_score":0.97,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":0.75,"output_cost_per_million":3.75,"speed_rps":null,"context_window":1048576,"release_date":"2026-08-13","announcement_date":"2026-08-13","multimodal":true,"param_count":null,"is_new":false},{"rank":2,"model_id":"qwen3.8-max","model_name":"Qwen3.8 Max","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.929,"normalized_score":0.929,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":1.65,"output_cost_per_million":4.951,"speed_rps":null,"context_window":256000,"release_date":"2026-08-02","announcement_date":"2026-08-02","multimodal":true,"param_count":2400000000000,"is_new":false},{"rank":3,"model_id":"gpt-5.6-sol","model_name":"GPT-5.6 Sol","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.915,"normalized_score":0.915,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":5.0,"output_cost_per_million":30.0,"speed_rps":null,"context_window":1050000,"release_date":"2026-07-09","announcement_date":"2026-07-09","multimodal":true,"param_count":null,"is_new":false},{"rank":4,"model_id":"gpt-5.6-terra","model_name":"GPT-5.6 Terra","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.896,"normalized_score":0.896,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":2.0,"output_cost_per_million":12.0,"speed_rps":null,"context_window":1050000,"release_date":"2026-07-09","announcement_date":"2026-07-09","multimodal":true,"param_count":null,"is_new":false},{"rank":5,"model_id":"claude-opus-4-6","model_name":"Claude Opus 4.6","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.76,"normalized_score":0.76,"verified":false,"self_reported":true,"provider_id":"anthropic","input_cost_per_million":5.0,"output_cost_per_million":25.0,"speed_rps":42.0,"context_window":1000000,"release_date":"2026-02-05","announcement_date":"2026-02-05","multimodal":true,"param_count":null,"is_new":false},{"rank":6,"model_id":"gpt-5.5","model_name":"GPT-5.5","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.74,"normalized_score":0.74,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":5.0,"output_cost_per_million":30.0,"speed_rps":null,"context_window":1050000,"release_date":"2026-04-23","announcement_date":"2026-04-23","multimodal":true,"param_count":null,"is_new":false},{"rank":7,"model_id":"gemma-4-31b-it","model_name":"Gemma 4 31B","organization_name":"Google","organization_id":"google","benchmark_score":0.664,"normalized_score":0.664,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.13,"output_cost_per_million":0.38,"speed_rps":null,"context_window":262144,"release_date":"2026-04-02","announcement_date":"2026-04-02","multimodal":true,"param_count":30700000000,"is_new":false},{"rank":8,"model_id":"gemini-3.1-flash-lite-preview","model_name":"Gemini 3.1 Flash-Lite","organization_name":"Google","organization_id":"google","benchmark_score":0.601,"normalized_score":0.601,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":0.25,"output_cost_per_million":1.5,"speed_rps":null,"context_window":1000000,"release_date":"2026-03-03","announcement_date":"2026-03-03","multimodal":true,"param_count":null,"is_new":false},{"rank":9,"model_id":"gemini-3.6-flash","model_name":"Gemini 3.6 Flash","organization_name":"Google","organization_id":"google","benchmark_score":0.54,"normalized_score":0.54,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":1.5,"output_cost_per_million":7.5,"speed_rps":null,"context_window":1048576,"release_date":"2026-07-21","announcement_date":"2026-07-21","multimodal":true,"param_count":null,"is_new":false},{"rank":10,"model_id":"gemma-4-26b-a4b-it","model_name":"Gemma 4 26B-A4B","organization_name":"Google","organization_id":"google","benchmark_score":0.441,"normalized_score":0.441,"verified":false,"self_reported":true,"provider_id":"novita","input_cost_per_million":0.13,"output_cost_per_million":0.4,"speed_rps":null,"context_window":262144,"release_date":"2026-04-02","announcement_date":"2026-04-02","multimodal":true,"param_count":25200000000,"is_new":false},{"rank":11,"model_id":"gemma-4-12b-it","model_name":"Gemma 4 12B","organization_name":"Google","organization_id":"google","benchmark_score":0.434,"normalized_score":0.434,"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":"2026-05-23","announcement_date":"2026-05-23","multimodal":true,"param_count":11959730224,"is_new":false},{"rank":12,"model_id":"gpt-5.6-luna","model_name":"GPT-5.6 Luna","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.413,"normalized_score":0.413,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":0.2,"output_cost_per_million":1.2,"speed_rps":null,"context_window":1050000,"release_date":"2026-07-09","announcement_date":"2026-07-09","multimodal":true,"param_count":null,"is_new":false},{"rank":13,"model_id":"gpt-5.4-mini","model_name":"GPT-5.4 mini","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.336,"normalized_score":0.336,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":0.75,"output_cost_per_million":4.5,"speed_rps":null,"context_window":400000,"release_date":"2026-03-17","announcement_date":"2026-03-17","multimodal":true,"param_count":null,"is_new":false},{"rank":14,"model_id":"gpt-5.4-nano","model_name":"GPT-5.4 nano","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.331,"normalized_score":0.331,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":0.2,"output_cost_per_million":1.25,"speed_rps":null,"context_window":400000,"release_date":"2026-03-17","announcement_date":"2026-03-17","multimodal":true,"param_count":null,"is_new":false},{"rank":15,"model_id":"gemini-3.5-flash","model_name":"Gemini 3.5 Flash","organization_name":"Google","organization_id":"google","benchmark_score":0.266,"normalized_score":0.266,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":1.5,"output_cost_per_million":9.0,"speed_rps":null,"context_window":1048576,"release_date":"2026-05-19","announcement_date":"2026-05-19","multimodal":true,"param_count":null,"is_new":false},{"rank":16,"model_id":"gemini-3.1-pro-preview","model_name":"Gemini 3.1 Pro","organization_name":"Google","organization_id":"google","benchmark_score":0.263,"normalized_score":0.263,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":2.5,"output_cost_per_million":15.0,"speed_rps":90.0,"context_window":1048576,"release_date":"2026-02-19","announcement_date":"2026-02-19","multimodal":true,"param_count":null,"is_new":false},{"rank":17,"model_id":"gemini-3-pro-preview","model_name":"Gemini 3 Pro","organization_name":"Google","organization_id":"google","benchmark_score":0.263,"normalized_score":0.263,"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":"2025-11-18","announcement_date":"2025-11-18","multimodal":true,"param_count":null,"is_new":false},{"rank":18,"model_id":"gemma-4-e4b-it","model_name":"Gemma 4 E4B","organization_name":"Google","organization_id":"google","benchmark_score":0.254,"normalized_score":0.254,"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":"2026-04-02","announcement_date":"2026-04-02","multimodal":true,"param_count":8000000000,"is_new":false},{"rank":19,"model_id":"gemini-3-flash-preview","model_name":"Gemini 3 Flash","organization_name":"Google","organization_id":"google","benchmark_score":0.221,"normalized_score":0.221,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":0.5,"output_cost_per_million":3.0,"speed_rps":null,"context_window":1000000,"release_date":"2025-12-17","announcement_date":"2025-12-17","multimodal":true,"param_count":null,"is_new":false},{"rank":20,"model_id":"gemini-3.5-flash-lite","model_name":"Gemini 3.5 Flash-Lite","organization_name":"Google","organization_id":"google","benchmark_score":0.213,"normalized_score":0.213,"verified":false,"self_reported":true,"provider_id":"google","input_cost_per_million":0.3,"output_cost_per_million":2.5,"speed_rps":null,"context_window":1048576,"release_date":"2026-07-21","announcement_date":"2026-07-21","multimodal":true,"param_count":null,"is_new":false},{"rank":21,"model_id":"gemma-4-e2b-it","model_name":"Gemma 4 E2B","organization_name":"Google","organization_id":"google","benchmark_score":0.191,"normalized_score":0.191,"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":"2026-04-02","announcement_date":"2026-04-02","multimodal":true,"param_count":5100000000,"is_new":false},{"rank":22,"model_id":"gemini-2.5-pro-preview-06-05","model_name":"Gemini 2.5 Pro Preview 06-05","organization_name":"Google","organization_id":"google","benchmark_score":0.164,"normalized_score":0.164,"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":"2025-06-05","announcement_date":"2025-06-05","multimodal":true,"param_count":null,"is_new":false},{"rank":23,"model_id":"gemma-3-27b-it","model_name":"Gemma 3 27B","organization_name":"Google","organization_id":"google","benchmark_score":0.135,"normalized_score":0.135,"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":"2025-03-12","announcement_date":"2025-03-12","multimodal":true,"param_count":27000000000,"is_new":false}]}