Devin Security adds Claude Mythos 5.2 and a $100 million credit program

Cognition added Anthropic's Claude Mythos 5.2 model to Devin Security and is offering more than $100 million in credits to eligible enterprise customers for vulnerability scanning.

Cognition announced on November 11, 2026, that it added Anthropic's Claude Mythos 5.2 cyber-focused model to Devin Security, its managed vulnerability-scanning service, according to the announcement on the Devin blog.

Concrete changes

Claude Mythos 5.2 is now available inside Devin Security, Cognition's managed vulnerability-scanning service.
Devin Security runs scans using frontier AI models, so security teams do not need separate model access or licensing.
Cognition is committing more than $100 million in Devin credits toward the effort.
Access is currently limited to eligible enterprise customers; other teams must apply through a contact form.
On an internal cyber benchmark covering 25 repositories across multiple languages and frameworks, Claude Mythos 5.2 scored 73.5, detecting 1,781 known vulnerabilities using an F2 precision/recall metric.

What it means for people building with AI

The move signals Cognition pushing Devin further into enterprise application security rather than coding alone, using a frontier model as a differentiator and subsidizing adoption through large credit grants. For teams evaluating agentic security tooling, it adds a vulnerability-scanning option backed by model-level benchmarking rather than vendor claims alone.

Availability and next steps

Cognition has not disclosed public per-seat pricing; scans are framed as free or credited for qualifying customers. The update follows a November 10, 2026 Devin Desktop and CLI update that added the Gemini 4 Carbon model.

Sources

Claude Mythos 5.2 is now available in Devin Security (Cognition AI, November 11, 2026, official)

Written with AI assistance, strictly from the sources listed above. Before it went live, automated checks verified its dates, sources and every language edition.