{"benchmark_id":"cyberseceval-4","benchmark_name":"CyberSecEval 4","benchmark_description":"CyberSecEval 4 is an evaluation suite covering cybersecurity-related capabilities and risks of large language models. The insecure-code-generation tracks measure whether a model produces vulnerable code: the Instruct track presents coding requests designed to elicit known insecure patterns, while the Autocomplete track prompts the model with code context leading up to a known insecure pattern, with vulnerabilities detected via static analysis.","max_score":1.0,"categories":["safety","code"],"modality":"text","total_models":1,"entries":[{"rank":1,"model_id":"mai-thinking-1","model_name":"MAI-Thinking-1","organization_name":"Microsoft","organization_id":"microsoft","benchmark_score":0.63,"normalized_score":0.63,"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":"2026-06-02","announcement_date":"2026-06-02","multimodal":false,"param_count":1000000000000,"is_new":false}]}