{"benchmark_id":"healthbench","benchmark_name":"HealthBench","benchmark_description":"An open-source benchmark for measuring performance and safety of large language models in healthcare, consisting of 5,000 multi-turn conversations evaluated by 262 physicians using 48,562 unique rubric criteria across health contexts and behavioral dimensions","max_score":1.0,"categories":["healthcare"],"modality":"text","total_models":14,"entries":[{"rank":1,"model_id":"claude-sonnet-5-5","model_name":"Claude Sonnet 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.654,"normalized_score":0.654,"verified":false,"self_reported":true,"provider_id":"anthropic","input_cost_per_million":2.0,"output_cost_per_million":10.0,"speed_rps":42.0,"context_window":1000000,"release_date":"2026-09-28","announcement_date":"2026-09-28","multimodal":true,"param_count":null,"is_new":false},{"rank":2,"model_id":"claude-haiku-5-5","model_name":"Claude Haiku 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.616,"normalized_score":0.616,"verified":false,"self_reported":true,"provider_id":"anthropic","input_cost_per_million":0.1,"output_cost_per_million":0.5,"speed_rps":null,"context_window":1000000,"release_date":"2026-10-07","announcement_date":"2026-10-07","multimodal":true,"param_count":null,"is_new":true},{"rank":3,"model_id":"claude-opus-5-5","model_name":"Claude Opus 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.606,"normalized_score":0.606,"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":4,"model_id":"qwen3.8-max","model_name":"Qwen3.8 Max","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.602,"normalized_score":0.602,"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":"gpt-6.1-sol","model_name":"GPT-6.1 Sol","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.585,"normalized_score":0.585,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":2.0,"output_cost_per_million":10.0,"speed_rps":null,"context_window":1050000,"release_date":"2026-09-29","announcement_date":"2026-09-29","multimodal":true,"param_count":null,"is_new":false},{"rank":6,"model_id":"kimi-k2-thinking-0905","model_name":"Kimi K2-Thinking-0905","organization_name":"Moonshot AI","organization_id":"moonshotai","benchmark_score":0.58,"normalized_score":0.58,"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":"2025-09-05","announcement_date":"2025-09-05","multimodal":false,"param_count":1000000000000,"is_new":false},{"rank":7,"model_id":"gpt-oss-120b","model_name":"GPT OSS 120B","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.576,"normalized_score":0.576,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.037,"output_cost_per_million":0.17,"speed_rps":null,"context_window":131072,"release_date":"2025-08-05","announcement_date":"2025-08-05","multimodal":false,"param_count":116800000000,"is_new":false},{"rank":8,"model_id":"gpt-5.6-terra","model_name":"GPT-5.6 Terra","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.57,"normalized_score":0.57,"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":9,"model_id":"gpt-5.6-sol","model_name":"GPT-5.6 Sol","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.57,"normalized_score":0.57,"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":10,"model_id":"gpt-5.6-luna","model_name":"GPT-5.6 Luna","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.558,"normalized_score":0.558,"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":11,"model_id":"gpt-5.3-chat-latest","model_name":"GPT-5.3 Chat","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.541,"normalized_score":0.541,"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":128000,"release_date":"2026-03-04","announcement_date":"2026-03-04","multimodal":true,"param_count":null,"is_new":false},{"rank":12,"model_id":"gpt-5.5-instant","model_name":"GPT-5.5 Instant","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.514,"normalized_score":0.514,"verified":false,"self_reported":true,"provider_id":"openai","input_cost_per_million":5.0,"output_cost_per_million":30.0,"speed_rps":100.0,"context_window":400000,"release_date":"2026-05-05","announcement_date":"2026-05-05","multimodal":true,"param_count":null,"is_new":false},{"rank":13,"model_id":"gpt-oss-20b","model_name":"GPT OSS 20B","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.425,"normalized_score":0.425,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.03,"output_cost_per_million":0.14,"speed_rps":null,"context_window":131072,"release_date":"2025-08-05","announcement_date":"2025-08-05","multimodal":false,"param_count":20900000000,"is_new":false},{"rank":14,"model_id":"seed-2.0-mini","model_name":"Seed 2.0 Mini","organization_name":"ByteDance","organization_id":"bytedance","benchmark_score":0.3,"normalized_score":0.3,"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}]}