Cisco Launches Antares SLMs to Tackle Cybersecurity Vulnerabilities at 172x Lower Cost Than GPT

Cisco unveils Antares, a series of small language models built specifically for cybersecurity vulnerability detection that match GPT-4.5 performance while running at 172 times lower cost, offering organizations a faster, more affordable defense against growing AI-driven cyber threats.

• Cisco unveils Antares, a series of small language models (SLMs) ranging from 350 million to 1 billion parameters, specifically designed to identify and triage known vulnerabilities in codebases, with a 3 billion parameter version coming soon.
• Benchmark testing shows Antares outperforms major models including Google Gemini, Qwen, and GPT-OSS on security tasks, matches GPT-5.5 performance, and runs at 172 times lower cost while operating locally on-device or on-premise servers.
• As cybersecurity defenders face growing backlogs and increasingly powerful AI-driven threats, Cisco's purpose-built SLMs are emerging as a cost-effective solution that processes vulnerability triage in seconds, giving organizations a faster and more affordable line of defense.