{"benchmark_id":"humanity's-last-exam-(with-tools,-text-only)","benchmark_name":"Humanity's Last Exam (with tools, text-only)","benchmark_description":"Text-only Humanity's Last Exam variant evaluated with tool use enabled.","max_score":1.0,"categories":["math","reasoning"],"modality":"text","total_models":10,"entries":[{"rank":1,"model_id":"claude-opus-5-5","model_name":"Claude Opus 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.677,"normalized_score":0.677,"verified":false,"self_reported":true,"provider_id":"anthropic","input_cost_per_million":4.0,"output_cost_per_million":20.0,"speed_rps":null,"context_window":1000000,"release_date":"2026-09-22","announcement_date":"2026-09-22","multimodal":true,"param_count":null,"is_new":false},{"rank":2,"model_id":"deepseek-v4-pro-0813","model_name":"DeepSeek-V4-Pro-0813","organization_name":"DeepSeek","organization_id":"deepseek","benchmark_score":0.6,"normalized_score":0.6,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":1.3,"output_cost_per_million":2.6,"speed_rps":null,"context_window":1048576,"release_date":"2026-08-13","announcement_date":"2026-08-13","multimodal":false,"param_count":1600000000000,"is_new":false},{"rank":3,"model_id":"gpt-6-astra","model_name":"GPT-6 Astra","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.572,"normalized_score":0.572,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":10.0,"output_cost_per_million":50.0,"speed_rps":null,"context_window":1050000,"release_date":"2026-09-04","announcement_date":"2026-09-03","multimodal":true,"param_count":null,"is_new":false},{"rank":4,"model_id":"qwen3.8-max","model_name":"Qwen3.8 Max","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.562,"normalized_score":0.562,"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":5,"model_id":"hy4-preview","model_name":"Hy4 preview","organization_name":"Tencent","organization_id":"tencent","benchmark_score":0.554,"normalized_score":0.554,"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":"2026-08-28","announcement_date":"2026-08-28","multimodal":false,"param_count":770000000000,"is_new":false},{"rank":6,"model_id":"hy3","model_name":"Hy3","organization_name":"Tencent","organization_id":"tencent","benchmark_score":0.532,"normalized_score":0.532,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.14,"output_cost_per_million":0.58,"speed_rps":null,"context_window":262144,"release_date":"2026-07-06","announcement_date":"2026-07-06","multimodal":false,"param_count":295000000000,"is_new":false},{"rank":7,"model_id":"qwen3-max-thinking","model_name":"Qwen3 Max Thinking","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.498,"normalized_score":0.498,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":1.2,"output_cost_per_million":6.0,"speed_rps":null,"context_window":256000,"release_date":"2026-02-13","announcement_date":"2026-02-13","multimodal":false,"param_count":1000000000000,"is_new":false},{"rank":8,"model_id":"step-3.7-flash","model_name":"Step 3.7 Flash","organization_name":"StepFun","organization_id":"stepfun","benchmark_score":0.472,"normalized_score":0.472,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.2,"output_cost_per_million":1.15,"speed_rps":null,"context_window":262144,"release_date":"2026-06-10","announcement_date":"2026-06-10","multimodal":true,"param_count":198000000000,"is_new":false},{"rank":9,"model_id":"inkling","model_name":"Inkling","organization_name":"Thinking Machines Lab","organization_id":"thinking-machines","benchmark_score":0.46,"normalized_score":0.46,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.95,"output_cost_per_million":4.05,"speed_rps":null,"context_window":524288,"release_date":"2026-07-21","announcement_date":"2026-07-21","multimodal":true,"param_count":975000000000,"is_new":false},{"rank":10,"model_id":"seed-1.8","model_name":"Seed 1.8","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.409,"normalized_score":0.409,"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}]}