{"benchmark_id":"mcp-universe","benchmark_name":"MCP-Universe","benchmark_description":"MCP-Universe evaluates LLMs on complex multi-step agentic tasks using Model Context Protocol (MCP) tools across diverse interactive environments, testing planning, tool orchestration, and task completion.","max_score":1.0,"categories":["agents","tool_calling"],"modality":"text","total_models":1,"entries":[{"rank":1,"model_id":"deepseek-v3.2","model_name":"DeepSeek-V3.2","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.459,"normalized_score":0.459,"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":"2025-12-01","announcement_date":"2025-12-01","multimodal":false,"param_count":685000000000,"is_new":false}]}