# Amodei says Anthropic's safety delay may have cost it an AI lead

DevFeed: [Amodei says Anthropic's safety delay may have cost it an AI lead](<https://devfeed.tech/articles/amodei-says-anthropic-s-safety-delay-may-have-cost-it-an-ai-lead-93943.md>)

Original publisher: [Read original article](<https://thenewstack.io/anthropic-mythos-safety-pace/>)

Author: Amanda Caswell

Published: 2026-10-21T21:22:53Z

Content type: news

Language: en

Sources: [The New Stack](<https://devfeed.tech/sources/the-new-stack.md>)

Topics: [anthropic](<https://devfeed.tech/topics/anthropic.md>), [ai safety](<https://devfeed.tech/topics/ai-safety.md>), [responsible-ai](<https://devfeed.tech/topics/responsible-ai.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [ai regulation](<https://devfeed.tech/topics/ai-regulation.md>), [Claude](<https://devfeed.tech/topics/claude.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [Vulnerabilities](<https://devfeed.tech/topics/vulnerabilities.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2023](<https://devfeed.tech/tags/2023.md>), [2026](<https://devfeed.tech/tags/2026.md>), [accelerate](<https://devfeed.tech/tags/accelerate.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [ai-strategy](<https://devfeed.tech/tags/ai-strategy.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [but](<https://devfeed.tech/tags/but.md>), [ceo](<https://devfeed.tech/tags/ceo.md>), [class](<https://devfeed.tech/tags/class.md>), [cost](<https://devfeed.tech/tags/cost.md>), [dario-amodei](<https://devfeed.tech/tags/dario-amodei.md>), [delay](<https://devfeed.tech/tags/delay.md>), [first](<https://devfeed.tech/tags/first.md>), [june](<https://devfeed.tech/tags/june.md>), [lead](<https://devfeed.tech/tags/lead.md>), [models](<https://devfeed.tech/tags/models.md>), [mythos](<https://devfeed.tech/tags/mythos.md>), [post](<https://devfeed.tech/tags/post.md>), [safety](<https://devfeed.tech/tags/safety.md>), [security](<https://devfeed.tech/tags/security.md>), [stack](<https://devfeed.tech/tags/stack.md>)

## AI overview

Anthropic CEO Dario Amodei says the company delayed broader access to Mythos-class models until June to give organizations time to prepare for cyber risks, conceding it may have sacrificed a several-month competitive lead. Anthropic's Responsible Scaling Policy now emphasizes transparency, risk reporting and independent evaluation; Amodei argues remaining at the frontier lets the company influence safety standards, while critics say policy changes weaken prior commitments. Access expanded through Project Glasswing and later programs, with different cybersecurity restrictions by model and partner tier. Anthropic's tests found its most restrictive tier blocked Opus 5.5 in 46 of 50 offensive-security trials, while a less restrictive tier blocked none; no comparable defensive benchmark is provided, and Anthropic acknowledges false positives on secure coding. The excerpt also describes later government export controls and their lifting, alongside planned cooperation on testing and cybersecurity standards.

## Source excerpt

Anthropic first released Mythos-class models in February 2026, but CEO Dario Amodei said it delayed broader access until June to The post Amodei says Anthropic's safety delay may have cost it an AI lead appeared first on The New Stack.