Cisco Foundation AI has introduced Antares, an open-weight small language model family designed to help security teams locate known vulnerabilities in codebases quickly and at lower cost. Cisco says the models, Antares-350M and Antares-1B, are especially useful for budget-constrained organizations and are backed by a new Vulnerability Localization Benchmark. #CiscoFoundationAI #Antares #VulnerabilityLocalizationBenchmark
Keypoints
- Antares is Cisco Foundation AI’s new open-weight small language model family.
- The models are built to locate known vulnerabilities inside codebases.
- Two versions are available: Antares-350M and Antares-1B.
- Cisco says Antares aims to reduce cost and false positives versus other AI options.
- The company also released the Vulnerability Localization Benchmark to measure performance.
Read More: https://www.securityweek.com/cisco-launches-low-cost-ai-models-for-source-code-security/