{"benchmark_id":"swe-bench-multimodal","benchmark_name":"SWE-Bench Multimodal","benchmark_description":"SWE-Bench Multimodal extends SWE-Bench to evaluate language models on software engineering tasks that involve visual inputs such as screenshots, UI mockups, and diagrams alongside code understanding.","max_score":1.0,"categories":["multimodal","reasoning","agents","code","vision"],"modality":"multimodal","total_models":4,"entries":[{"rank":1,"model_id":"claude-mythos-preview","model_name":"Claude Mythos Preview","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.59,"normalized_score":0.59,"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":null,"announcement_date":"2026-04-07","multimodal":true,"param_count":null,"is_new":false},{"rank":2,"model_id":"qwen3.8-27b","model_name":"Qwen3.8-27B","organization_name":"Alibaba Cloud / Qwen Team","organization_id":"qwen","benchmark_score":0.386,"normalized_score":0.386,"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":"2026-08-14","announcement_date":"2026-08-14","multimodal":true,"param_count":27781427952,"is_new":false},{"rank":3,"model_id":"claude-opus-4-8","model_name":"Claude Opus 4.8","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.384,"normalized_score":0.384,"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":4,"model_id":"claude-sonnet-5","model_name":"Claude Sonnet 5","organization_name":"Anthropic","organization_id":"anthropic","benchmark_score":0.281,"normalized_score":0.281,"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":"2026-06-30","announcement_date":"2026-06-30","multimodal":true,"param_count":null,"is_new":false}]}