Cisco releases Antares models for local vulnerability localization in code
aminkarbasi · x · 2026-07-22
Cisco introduced Antares, a family of small language models for locating known vulnerabilities in code.
- Two models are live on Hugging Face: Antares-350M and Antares-1B.
- Cisco says they can outperform many larger closed- and open-weight models at a fraction of the cost.
- The models are compact enough to run locally, so teams do not need to ship sensitive codebases to the cloud.
- Cisco also released a new Vulnerability Localization Benchmark to support the work.
The pitch is straightforward: vulnerability triage is expensive and slow, and smaller open-weight models may make AI-assisted security more practical.
More from Safety
- OpenAI test model reportedly broke containment and hit a real company’s servers — cnn · 2026-07-22
- OpenAI says its AI went rogue and launched an unprecedented cyberattack — KeanuRave100 · 2026-07-22
- AI Security Institute tests lie detectors across 31 open-weight models — geoffreyirving · 2026-07-22
- Australia Gears Up for New AI Rules, Impacting OpenAI and Anthropic — nordicinst · 2026-07-22
- AI access is outpacing operational control, and agents need workflow-level permissions — Early-Matter-8123 · 2026-07-22
- Telemetry can’t prove an AI intrusion was fully autonomous — cyb3rops · 2026-07-22