Osegi
AI·2 min read

Anthropic's Mythos Model Comes to Cisco and IBM Security Tools

Cisco and IBM will build Anthropic's bug-finding Mythos into AI Defense and Concert, giving customers its capabilities without access to the model itself.

Abstract illustration of an AI model scanning enterprise software for security vulnerabilities, with network security icons.
Image: theregister
SUMODialoediviewergiftdetect

Anthropic is widening access to its bug-finding Mythos model: Cisco and IBM will both build it into their security products, letting customers use its capabilities without getting the model itself.

Cisco announced its integration at an event in Melbourne, Australia. The red teaming tools in its AI Defense suite, which probe AI models, agents and the applications built on them for weaknesses and inappropriate behavior, will be able to call Mythos to test customers' setups. "At a very high level, AI Defense has the ability to look at models and break them," said Jeff Schultz, Cisco's senior vice president for portfolio strategy and marketing, adding that the tool is now "super-charged by Mythos." AI Defense already uses AI-powered red teaming, and Mythos becomes another option. Cisco will offer it under existing licenses at no extra cost.

Hours later, IBM said it will integrate Mythos into Concert, its tool that uses AI to scan applications for risks. IBM says the model will help "uncover weaknesses traditional scanners miss, determine which pose real risk and generate fixes for developer approval."

Controlled access, growing demand

Anthropic first released Mythos through Project Glasswing, a program giving vetted organizations access to test their codebases, and has since expanded through other controlled programs. The caution is deliberate: in the wrong hands, Mythos could help attackers exploit software flaws. The Cisco and IBM approach exposes the functionality to end users while keeping the model out of their hands.

Natalie Mead, Anthropic's head of applied AI for Asia-Pacific, said the company has a long queue of buyers wanting Mythos and that working with Cisco makes it easier to meet that demand. She declined to say whether similar deals exist with other third parties.

Version 5.2 arrives

Also on Monday, Anthropic introduced version 5.2 of Mythos and Fable, which it says generate output up to 30 percent faster than 5.1 at lower prices: $8 per million input tokens and $40 per million output tokens. A "fast mode" speeds up generation by up to 2.5 times and doubles prices to $16 and $80.

Anthropic says prerelease testing concluded that high-stakes misalignment risks remain low. It adds that Fable 5.2's cyber and biology safeguards are as robust as 5.1's with fewer unnecessary blocks, and that it has put in place its strongest-ever protections against distillation attacks.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:anthropicmythosciscoibmai securityred teaming

More from this day

See all →

More in: AI

See all →