{"benchmark_id":"mcp-atlas","benchmark_name":"MCP Atlas","benchmark_description":"MCP Atlas is a benchmark for evaluating AI models on scaled tool use capabilities, measuring how well models can coordinate and utilize multiple tools across complex multi-step tasks.","max_score":1.0,"categories":["reasoning","agents","code","tool_calling"],"modality":"text","total_models":33,"entries":[{"rank":1,"model_id":"muse-spark-1.1","model_name":"Muse Spark 1.1","organization_name":"Meta","organization_id":"meta","benchmark_score":0.881,"normalized_score":0.881,"verified":false,"self_reported":true,"provider_id":"meta","input_cost_per_million":1.25,"output_cost_per_million":4.25,"speed_rps":null,"context_window":1048576,"release_date":"2026-07-09","announcement_date":"2026-07-09","multimodal":true,"param_count":null,"is_new":false},{"rank":2,"model_id":"kimi-k3","model_name":"Kimi K3","organization_name":"Moonshot AI","organization_id":"moonshotai","benchmark_score":0.842,"normalized_score":0.842,"verified":false,"self_reported":true,"provider_id":"novita","input_cost_per_million":3.0,"output_cost_per_million":15.0,"speed_rps":null,"context_window":1048576,"release_date":"2026-07-16","announcement_date":"2026-07-16","multimodal":true,"param_count":2800000000000,"is_new":false},{"rank":3,"model_id":"seed-2.1-pro","model_name":"Seed 2.1 Pro","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.838,"normalized_score":0.838,"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":"2026-06-24","announcement_date":"2026-06-24","multimodal":true,"param_count":null,"is_new":false},{"rank":4,"model_id":"gemini-3.5-flash","model_name":"Gemini 3.5 Flash","organization_name":"Google","organization_id":"google","benchmark_score":0.836,"normalized_score":0.836,"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":5,"model_id":"claude-opus-4-8","model_name":"Claude Opus 4.8","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.822,"normalized_score":0.822,"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":"2026-05-28","announcement_date":"2026-05-28","multimodal":true,"param_count":null,"is_new":false},{"rank":6,"model_id":"seed-2.1-turbo","model_name":"Seed 2.1 Turbo","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.803,"normalized_score":0.803,"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":"2026-06-24","announcement_date":"2026-06-24","multimodal":true,"param_count":null,"is_new":false},{"rank":7,"model_id":"inkling-small","model_name":"Inkling-Small","organization_name":"Thinking Machines Lab","organization_id":"thinking-machines","benchmark_score":0.796,"normalized_score":0.796,"verified":false,"self_reported":true,"provider_id":"thinking-machines","input_cost_per_million":0.3,"output_cost_per_million":1.2,"speed_rps":null,"context_window":256000,"release_date":"2026-07-30","announcement_date":"2026-07-30","multimodal":true,"param_count":276000000000,"is_new":false},{"rank":8,"model_id":"hy3","model_name":"Hy3","organization_name":"Tencent","organization_id":"tencent","benchmark_score":0.791,"normalized_score":0.791,"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":"2026-07-06","announcement_date":"2026-07-06","multimodal":false,"param_count":295000000000,"is_new":false},{"rank":9,"model_id":"claude-opus-4-7","model_name":"Claude Opus 4.7","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.773,"normalized_score":0.773,"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":"2026-04-16","announcement_date":"2026-04-16","multimodal":true,"param_count":null,"is_new":false},{"rank":10,"model_id":"glm-5.2","model_name":"GLM-5.2","organization_name":"Zhipu AI","organization_id":"zai-org","benchmark_score":0.768,"normalized_score":0.768,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.95,"output_cost_per_million":3.0,"speed_rps":null,"context_window":1048576,"release_date":"2026-06-16","announcement_date":"2026-06-16","multimodal":false,"param_count":753000000000,"is_new":false},{"rank":11,"model_id":"qwen3.7-max","model_name":"Qwen3.7 Max","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.764,"normalized_score":0.764,"verified":false,"self_reported":true,"provider_id":"novita","input_cost_per_million":1.25,"output_cost_per_million":3.75,"speed_rps":null,"context_window":1000000,"release_date":"2026-05-19","announcement_date":"2026-05-19","multimodal":false,"param_count":null,"is_new":false},{"rank":12,"model_id":"kimi-k2.7-code","model_name":"Kimi K2.7 Code","organization_name":"Moonshot AI","organization_id":"moonshotai","benchmark_score":0.76,"normalized_score":0.76,"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":"2026-06-12","announcement_date":"2026-06-12","multimodal":true,"param_count":1000000000000,"is_new":false},{"rank":13,"model_id":"muse-glimmer-30b","model_name":"Muse Glimmer-30B","organization_name":"Meta","organization_id":"meta","benchmark_score":0.755,"normalized_score":0.755,"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":"2026-08-10","announcement_date":"2026-08-10","multimodal":true,"param_count":29600000000,"is_new":false},{"rank":14,"model_id":"gpt-5.5","model_name":"GPT-5.5","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.753,"normalized_score":0.753,"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":15,"model_id":"minimax-m3","model_name":"MiniMax M3","organization_name":"MiniMax","organization_id":"minimax","benchmark_score":0.742,"normalized_score":0.742,"verified":false,"self_reported":true,"provider_id":"novita","input_cost_per_million":0.3,"output_cost_per_million":1.2,"speed_rps":null,"context_window":1000000,"release_date":"2026-06-01","announcement_date":"2026-06-01","multimodal":true,"param_count":428000000000,"is_new":false},{"rank":16,"model_id":"qwen3.6-plus","model_name":"Qwen3.6 Plus","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.741,"normalized_score":0.741,"verified":false,"self_reported":true,"provider_id":"together","input_cost_per_million":0.5,"output_cost_per_million":3.0,"speed_rps":null,"context_window":1000000,"release_date":"2026-03-31","announcement_date":"2026-04-02","multimodal":true,"param_count":null,"is_new":false},{"rank":17,"model_id":"deepseek-v4-pro-max","model_name":"DeepSeek-V4-Pro-Max","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.736,"normalized_score":0.736,"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":18,"model_id":"qwen3.7-plus","model_name":"Qwen3.7-Plus","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.732,"normalized_score":0.732,"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":19,"model_id":"glm-5.1","model_name":"GLM-5.1","organization_name":"Zhipu AI","organization_id":"zai-org","benchmark_score":0.718,"normalized_score":0.718,"verified":false,"self_reported":true,"provider_id":"friendli","input_cost_per_million":1.4,"output_cost_per_million":4.4,"speed_rps":null,"context_window":200000,"release_date":"2026-04-07","announcement_date":"2026-04-07","multimodal":false,"param_count":754000000000,"is_new":false},{"rank":20,"model_id":"gemini-3.1-pro-preview","model_name":"Gemini 3.1 Pro","organization_name":"Google","organization_id":"google","benchmark_score":0.692,"normalized_score":0.692,"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":21,"model_id":"deepseek-v4-flash-max","model_name":"DeepSeek-V4-Flash-Max","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.69,"normalized_score":0.69,"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":22,"model_id":"glm-5","model_name":"GLM-5","organization_name":"Zhipu AI","organization_id":"zai-org","benchmark_score":0.678,"normalized_score":0.678,"verified":false,"self_reported":true,"provider_id":"friendli","input_cost_per_million":1.0,"output_cost_per_million":3.2,"speed_rps":null,"context_window":200000,"release_date":"2026-02-11","announcement_date":"2026-02-11","multimodal":false,"param_count":744000000000,"is_new":false},{"rank":23,"model_id":"deepseek-v4-flash-0423","model_name":"DeepSeek-V4-Flash-0423","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.674,"normalized_score":0.674,"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":24,"model_id":"gpt-5.4","model_name":"GPT-5.4","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.672,"normalized_score":0.672,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":2.5,"output_cost_per_million":15.0,"speed_rps":50.0,"context_window":1000000,"release_date":"2026-03-05","announcement_date":"2026-03-05","multimodal":true,"param_count":null,"is_new":false},{"rank":25,"model_id":"qwen3.6-35b-a3b","model_name":"Qwen3.6-35B-A3B","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.628,"normalized_score":0.628,"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":"2026-04-16","announcement_date":"2026-04-16","multimodal":true,"param_count":35000000000,"is_new":false},{"rank":26,"model_id":"claude-opus-4-6","model_name":"Claude Opus 4.6","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.627,"normalized_score":0.627,"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":27,"model_id":"claude-opus-4-5-20251101","model_name":"Claude Opus 4.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.623,"normalized_score":0.623,"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":"2025-11-24","announcement_date":"2025-11-24","multimodal":true,"param_count":null,"is_new":false},{"rank":28,"model_id":"claude-sonnet-4-6","model_name":"Claude Sonnet 4.6","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.613,"normalized_score":0.613,"verified":false,"self_reported":true,"provider_id":"anthropic","input_cost_per_million":3.0,"output_cost_per_million":15.0,"speed_rps":42.0,"context_window":200000,"release_date":"2026-02-17","announcement_date":"2026-02-17","multimodal":true,"param_count":null,"is_new":false},{"rank":29,"model_id":"gpt-5.2-2025-12-11","model_name":"GPT-5.2","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.606,"normalized_score":0.606,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":1.75,"output_cost_per_million":14.0,"speed_rps":100.0,"context_window":400000,"release_date":"2025-12-11","announcement_date":"2025-12-11","multimodal":true,"param_count":null,"is_new":false},{"rank":30,"model_id":"gpt-5.4-mini","model_name":"GPT-5.4 mini","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.577,"normalized_score":0.577,"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":31,"model_id":"gemini-3-flash-preview","model_name":"Gemini 3 Flash","organization_name":"Google","organization_id":"google","benchmark_score":0.574,"normalized_score":0.574,"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":32,"model_id":"gpt-5.4-nano","model_name":"GPT-5.4 nano","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.561,"normalized_score":0.561,"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":33,"model_id":"nova-2-lite","model_name":"Nova 2 Lite","organization_name":"Amazon","organization_id":"amazon","benchmark_score":0.246,"normalized_score":0.246,"verified":false,"self_reported":true,"provider_id":"bedrock","input_cost_per_million":0.3,"output_cost_per_million":2.5,"speed_rps":null,"context_window":1000000,"release_date":"2025-12-02","announcement_date":"2025-12-02","multimodal":true,"param_count":null,"is_new":false}]}