This week, Cisco introduced Antares — a family of open-weight small language models designed to help localize known vulnerabilities inside an organization’s own infrastructure, at a fraction of the cost of running that work through a frontier model.
For partners, the opportunity starts with what they can build around it. Because Antares is being released as an open-weight capability, partners can use it as a foundation for services like assessments, advisory work, managed workflows, and AI-enabled vulnerability operations that help customers move from finding issues to acting on them.
Why this matters to customers
According to Talos, last year more than 20% of the most targeted vulnerabilities were over a decade old. Defenders usually know these flaws exist but cannot get to all of them, everywhere, fast enough.
Most security teams have looked to AI to help, but fewer have deployed it. In some cases, it’s because they cannot send source code outside their own environment. In others, it’s because running large models against a codebase gets so expensive that scanning becomes selective.
Antares addresses both challenges. Code never leaves the environment, and based on benchmark testing, Antares-350M and Antares-1B outperform many powerful closed- and open-weight models in this critical security task at a fraction of the cost. and stops forcing teams to choose which codebases matter most.
This is not a replacement for frontier models. Antares takes a narrower focus: finding where an already-known vulnerability lives in your code. Put another way, using frontier AI to hunt known flaws in your own code is a bit like taking a private jet to the corner store. It works, but you wouldn’t do it for every trip.
Expanding access and action
That’s where partners play a critical role.
In practice, this could look like a partner-led vulnerability assessment, secure code review service, remediation planning engagement, or ongoing managed workflow that helps customers continuously triage known vulnerabilities across large code environments.
The second opportunity is about access. Local deployment opens conversations with customers who have been cautious about AI security because cloud was never going to clear their compliance or budget requirements. Public sector, universities, smaller teams, regulated enterprises are all examples of where Antares can create a more practical path forward.
Building toward ongoing vulnerability operations
Right now, every customer is asking some version of the same question: am I vulnerable? Historically, the answer came from a point-in-time snapshot and assessment.
Antares makes a different motion possible. It gives partners a building block for ongoing vulnerability operations, helping customers detect, prioritize, and respond as things change rather than waiting for the next scheduled review.
For partners running managed practices, this is a natural extension of work they already do for customers. For others, it creates room for professional services around assessments, triage, remediation support, and secure AI adoption.
Either way, security is moving toward continuous operations – and the partners building for it now are establishing the kind of durable, outcome-oriented work that carries weight across the Cisco partner ecosystem.
Visit our Hugging Face page to explore the Antares models and review the model card. To learn more about the benchmark and evaluation methodology, read the technical paper, or contact Cisco Foundation AI.