Articles liés
Cyber Index
Benchmark evaluating a model's ability to reproduce real vulnerabilities in open-source software and subsequently patch them; Mistral Large 4 scored 82%.
Articles liés
Benchmark evaluating a model's ability to reproduce real vulnerabilities in open-source software and subsequently patch them; Mistral Large 4 scored 82%.