{"benchmark_id":"benchcad-with-python-tool","benchmark_name":"BenchCAD (with Python tool)","benchmark_description":"BenchCAD variant evaluated with access to a Python tool for programmatic CAD reasoning.","max_score":1.0,"categories":["multimodal","reasoning","code","vision"],"modality":"multimodal","total_models":7,"entries":[{"rank":1,"model_id":"claude-sonnet-5-5","model_name":"Claude Sonnet 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.963,"normalized_score":0.963,"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-opus-5-5","model_name":"Claude Opus 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.962,"normalized_score":0.962,"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":3,"model_id":"gpt-6-astra","model_name":"GPT-6 Astra","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.959,"normalized_score":0.959,"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":"claude-haiku-5-5","model_name":"Claude Haiku 5.5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.87,"normalized_score":0.87,"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":5,"model_id":"gpt-5.6-sol","model_name":"GPT-5.6 Sol","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.834,"normalized_score":0.834,"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":6,"model_id":"gpt-5.6-terra","model_name":"GPT-5.6 Terra","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.782,"normalized_score":0.782,"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":7,"model_id":"gpt-5.6-luna","model_name":"GPT-5.6 Luna","organization_name":"OpenAI","organization_id":"openai","benchmark_score":0.739,"normalized_score":0.739,"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}]}