Cognition upgrades Devin Security with Claude Mythos deep scans and pledges hundreds of millions of dollars in free credits
We are entering a new dark age of cybersecurity. Devin is here to help. Introducing the new Devin Security: powered by Claude Mythos and frontier cyber models, plus hundreds of millions of dollars of credits to help teams find, prioritize, and fix urgent vulnerabilities.
Context
It gives enterprise security teams access to Anthropic's restricted Claude Mythos cyber model through Devin without their own model access, and pairs it with a large credit program. The credit total and the product's performance are Cognition's own claims, with no per customer amounts or benchmarks published.
Cognition makes the Devin coding agent, and Devin Security is its product for finding, prioritizing and fixing software vulnerabilities. Anthropic has limited access to its Claude Mythos class of cyber capable models to selected organizations. Cognition announced the upgrade on October 26 at 11:30 AM Eastern.
Deep dive
Cognition launched a major upgrade to Devin Security on October 26. According to the company's blog post, Devin Security now offers deep scans powered by Claude Mythos and other frontier cyber models, and enterprise security teams can use those capabilities through Devin without needing separate access to the models.
Cognition says deep scans are meant to uncover complex chained exploits across large codebases, guided by a custom threat model built with each customer. The product scores findings by how exploitable and damaging they would be in that environment, then has Devin write, validate and test fixes on real virtual machines in parallel. It can also ingest findings from existing scanners, bug bounties and advisories, remove duplicates and recheck them with frontier models. Cognition describes the product as model neutral, able to switch to whichever cyber capable models work best.
The second part of the launch is money. Cognition says it is committing hundreds of millions of dollars in Devin credits, offering current and future enterprise customers free scans with frontier cyber models and remediation credits for new customers. The company has not published a per customer amount, eligibility rules, pricing or benchmarks for the Mythos powered deep scans, as outside coverage from explainx and The Hype pointed out.
Cognition's launch post framed the moment as a new dark age of cybersecurity, a reference to how frontier models are making attacks faster and cheaper.
What to expect next
Watch for the first customer results and any published benchmark for deep scans, since Cognition has so far released none for the Mythos version. The credit program's terms, including caps and expiry, will show how much of the headline figure reaches security teams. Anthropic's own access program for Mythos, and whether other agent companies resell restricted cyber models the same way, will shape how widely these tools spread.