{"benchmark_id":"pinchbench","benchmark_name":"PinchBench","benchmark_description":"PinchBench evaluates coding agents on real-world agentic coding tasks, measuring both best-case and average performance across complex software engineering scenarios.","max_score":1.0,"categories":["agents","code"],"modality":"text","total_models":6,"entries":[{"rank":1,"model_id":"nemotron-3-ultra-550b-a55b","model_name":"Nemotron 3 Ultra (550B A55B)","organization_name":"NVIDIA","organization_id":"nvidia","benchmark_score":0.9,"normalized_score":0.9,"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":"2026-06-04","announcement_date":"2026-06-04","multimodal":false,"param_count":550000000000,"is_new":false},{"rank":2,"model_id":"nemotron-3.5-lightning-30b-a3b","model_name":"Nemotron 3.5 Lightning (30B A3B)","organization_name":"NVIDIA","organization_id":"nvidia","benchmark_score":0.8537,"normalized_score":0.8537,"verified":false,"self_reported":true,"provider_id":"deepinfra","input_cost_per_million":0.05,"output_cost_per_million":0.2,"speed_rps":null,"context_window":262144,"release_date":"2026-08-11","announcement_date":"2026-08-11","multimodal":false,"param_count":30000000000,"is_new":false},{"rank":3,"model_id":"mimo-v2-omni","model_name":"MiMo-V2-Omni","organization_name":"Xiaomi","organization_id":"xiaomi","benchmark_score":0.812,"normalized_score":0.812,"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":"2026-03-18","announcement_date":"2026-03-18","multimodal":true,"param_count":null,"is_new":false},{"rank":4,"model_id":"mimo-v2-pro","model_name":"MiMo-V2-Pro","organization_name":"Xiaomi","organization_id":"xiaomi","benchmark_score":0.81,"normalized_score":0.81,"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":"2026-03-18","announcement_date":"2026-03-18","multimodal":false,"param_count":1000000000000,"is_new":false},{"rank":5,"model_id":"glm-5v-turbo","model_name":"GLM-5V-Turbo","organization_name":"Zhipu AI","organization_id":"zai-org","benchmark_score":0.807,"normalized_score":0.807,"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":"2026-04-02","announcement_date":"2026-04-02","multimodal":true,"param_count":null,"is_new":false},{"rank":6,"model_id":"lfm-2.5-2.6b","model_name":"LFM2.5-2.6B","organization_name":"Liquid AI","organization_id":"liquid","benchmark_score":0.6822,"normalized_score":0.6822,"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":"2026-08-04","announcement_date":"2026-08-04","multimodal":false,"param_count":2690000000,"is_new":false}]}