Anthropic deploys Claude Mythos 5 to security scans, launches $35M fund

Anthropic is pushing frontier AI into enterprise security, deploying Claude Mythos 5 to power Claude Security Scans and committing $35 million to open-source network defense.

The deployment follows federal researchers' finding that Anthropic's Fable 5 model could be prompted to patch vulnerable code without a jailbreak, a result that shows both the capability and the risk of handing security work to frontier AI.

Claude Security Scans now runs on Claude Mythos 5, the newest member of Anthropic's model family that also includes Opus 5 and Fable 5. The $35 million fund is earmarked for open-source network security projects, part of a broader push to embed AI into the tools security teams already use.

The move puts Anthropic in direct competition with OpenAI, which has expanded its own agentic security tools, and with dedicated security vendors such as CrowdStrike and Palo Alto Networks that are racing to add AI to their platforms.

The security push comes as AI vendors race to automate the most labor-intensive parts of security operations. Security Operations Centers receive a high volume of alerts from endpoints, cloud workloads, network devices, identity providers, and business applications, and analysts often spend considerable time correlating alerts and determining next investigative steps. AI offers a practical way to augment analysts by providing contextual explanations and summarizing findings.

Anthropic's Claude Security Scans product is designed to automate code review and vulnerability detection. Running it on Mythos 5, the company's latest model, is intended to improve the accuracy and speed of those scans. Anthropic did not disclose benchmark results for Mythos 5 on security tasks, nor the model's parameter count.

The $35 million fund is a bet on the open-source security community, which maintains much of the critical infrastructure software that enterprises rely on. Funding open-source projects has become a common strategy for large technology companies seeking to harden the software supply chain, following similar commitments from rivals that have poured money into the same projects.

The competitive stakes

The enterprise security market is a large and growing prize. AI vendors see security as one of the most commercially promising applications of their models, because the cost of a breach is high and the demand for automation is strong. OpenAI has been expanding its own security offerings, and Anthropic's move is a direct challenge to that effort.

For investors, the question is whether AI security tools can translate into durable revenue. Anthropic has not disclosed pricing for Claude Security Scans or the expected revenue contribution. The company's broader model lineup — Opus 5, Fable 5, and now Mythos 5 — suggests it is betting that specialized models for specific tasks will win over general-purpose assistants.

The deployment is a reminder that the AI security market is becoming crowded. Dedicated security vendors such as CrowdStrike and Palo Alto Networks are adding AI features to their platforms, while AI labs such as Anthropic and OpenAI are moving into the space from the other direction. Which approach wins will depend on whether enterprises trust frontier AI models with their most sensitive security workloads.

This article is for informational purposes only and does not constitute investment advice.