CISCO LAUNCHES AI MODELS FOR CODE VULNERABILITY DETECTION

AI DESK■ 1 MIN READ

TUE, JUL 21, 2026

■ AI-SUMMARIZED FROM 2 SOURCES ▸ TIMELINE

Cisco released two open-weight AI models, Antares-350M and Antares-1B, designed to identify known vulnerabilities in codebases. The company plans to release a larger Antares-3B model soon.

The models perform comparably to significantly larger language models, including OpenAI's GPT-5.5 and Z.ai's GLM-5.2, according to Cisco. Antares represents Cisco's effort to democratize security tooling through open-source AI. By releasing models of varying sizes, the company enables developers to run vulnerability detection locally without reliance on proprietary solutions. The open-weight approach allows organizations to fine-tune models for their specific codebases and security requirements. This contrasts with closed-source alternatives that limit customization and transparency. The roadmap includes Antares-3B, which will offer additional capability for security teams managing larger codebases. Cisco's release aligns with broader industry trends toward open-source AI models and developer-focused security tools. No pricing or exact release date for Antares-3B was disclosed.

■ SOURCES

► Techmeme ► The Decoder

■ SUMMARY WRITTEN BY AI FROM THE LINKS ABOVE

■ MORE FROM THE SECURITY DESK

P542N-ABLE PATCHES CRITICAL N-CENTRAL FLAW

N-able has released an emergency hotfix for a maximum-severity remote code execution vulnerability in its N-central RMM platform. The flaw is being actively exploited in ongoing attacks.

JUST NOW— Industry Desk

P519QBITTORRENT ESCAPES SANDBOX IN SECURITY BREACH

QBittorrent, the popular open-source torrent client, has been found capable of breaking out of sandbox environments to execute unauthorized operations. Security researchers identified the vulnerability, raising concerns about the application's access to system resources.

14H AGO— Industry Desk

P518PHISHING ATTACKS USE HIDDEN UNICODE TO BYPASS EMAIL FILTERS

Threat actors are deploying invisible Unicode characters in phishing campaigns to evade email security systems. The ASCII smuggling technique allows attackers to conceal malicious content from detection tools.

15H AGO— Security Desk

P511ASCII SMUGGLING SHIFTS FROM AI DEFENSE TO SPAM TOOL

A Unicode block invisible to human readers has transitioned from an academic curiosity used to test AI systems into an active tool for spammers. The technique exploits characters that machines process but humans cannot see.

20H AGO— AI Desk

◄ BACK TO NEWS

CISCO LAUNCHES AI MODELS FOR CODE VULNERABILITY DETECTION

■ MORE FROM THE SECURITY DESK

■ SUBSCRIBE TO THE DAILY BRIEF