Cisco Foundation AI Releases Antares: 350M and 1B Open-Weight Models That Localize Known Vulnerabilities Inside Real Codebases

Cisco Foundation AI has introduced Antares, a collection of small open-weight models designed to identify known security vulnerabilities within software code. Available in 350M and 1B parameter sizes, these models perform competitively against significantly larger systems on the Vulnerability Localization Benchmark. By focusing on localizing weaknesses, the tool provides developers with a more efficient way to scan codebases without requiring the high computational resources needed by massive language models.