{"benchmark_id":"mmvu","benchmark_name":"MMVU","benchmark_description":"MMVU (Multimodal Multi-disciplinary Video Understanding) is a benchmark for evaluating multimodal models on video understanding tasks across multiple disciplines, testing comprehension and reasoning capabilities on video content.","max_score":1.0,"categories":["multimodal","reasoning","video","vision"],"modality":"multimodal","total_models":7,"entries":[{"rank":1,"model_id":"glm-5.3-flash","model_name":"GLM-5.3-Flash","organization_name":"Zhipu AI","organization_id":"zai-org","benchmark_score":0.805,"normalized_score":0.805,"verified":false,"self_reported":true,"provider_id":"z","input_cost_per_million":0.15,"output_cost_per_million":0.5,"speed_rps":null,"context_window":1000000,"release_date":"2026-08-26","announcement_date":"2026-08-26","multimodal":true,"param_count":320000000000,"is_new":false},{"rank":2,"model_id":"kimi-k2.5","model_name":"Kimi K2.5","organization_name":"Moonshot AI","organization_id":"moonshotai","benchmark_score":0.804,"normalized_score":0.804,"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":"2026-01-27","announcement_date":"2026-01-27","multimodal":true,"param_count":1000000000000,"is_new":false},{"rank":3,"model_id":"qwen3.5-122b-a10b","model_name":"Qwen3.5-122B-A10B","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.747,"normalized_score":0.747,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.29,"output_cost_per_million":2.4,"speed_rps":null,"context_window":262144,"release_date":"2026-02-24","announcement_date":"2026-02-24","multimodal":true,"param_count":122000000000,"is_new":false},{"rank":4,"model_id":"qwen3.5-27b","model_name":"Qwen3.5-27B","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.733,"normalized_score":0.733,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.26,"output_cost_per_million":2.6,"speed_rps":null,"context_window":262144,"release_date":"2026-02-24","announcement_date":"2026-02-24","multimodal":true,"param_count":27000000000,"is_new":false},{"rank":5,"model_id":"seed-1.8","model_name":"Seed 1.8","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.731,"normalized_score":0.731,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.25,"output_cost_per_million":2.0,"speed_rps":null,"context_window":256000,"release_date":"2026-02-17","announcement_date":"2026-02-17","multimodal":true,"param_count":null,"is_new":false},{"rank":6,"model_id":"qwen3.5-35b-a3b","model_name":"Qwen3.5-35B-A3B","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.723,"normalized_score":0.723,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.14,"output_cost_per_million":1.0,"speed_rps":null,"context_window":262144,"release_date":"2026-02-24","announcement_date":"2026-02-24","multimodal":true,"param_count":35000000000,"is_new":false},{"rank":7,"model_id":"seed-2.0-mini","model_name":"Seed 2.0 Mini","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.69,"normalized_score":0.69,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.1,"output_cost_per_million":0.4,"speed_rps":null,"context_window":256000,"release_date":"2026-02-17","announcement_date":"2026-02-17","multimodal":true,"param_count":null,"is_new":false}]}