Anthropic is releasing its most powerful and controversial AI model to the public today — June 10th 2026. Known internally as Claude Mythos and launching publicly under the name Claude Fable, this is the model Anthropic previously said was too dangerous to release publicly. Here’s everything you need to know.
What is Claude Mythos?
Claude Mythos is a large language model developed by Anthropic to find software vulnerabilities, used by a consortium of companies to secure software systems. Anthropic had not released the model to the public, citing safety and misuse concerns.
The model first became publicly known in March 2026 when a security misconfiguration at Anthropic accidentally exposed draft blog posts and internal documents revealing its existence. The leak confirmed what many in the security industry had suspected — Anthropic had built something significantly more powerful than anything previously available.
Project Glasswing — The Restricted Preview
Rather than releasing Claude Mythos directly to the public Anthropic took an unusual approach. Anthropic provided access to the Claude Mythos Preview to 50 initial partners including US federal agencies through Project Glasswing — a restricted initiative to help organisations find vulnerabilities and prepare response measures using the AI model.
Anthropic announced it would expand the Glasswing project in early June to 150 organisations in more than 15 countries.
Partners in the restricted preview included AWS, Microsoft, Apple and CrowdStrike — some of the most significant names in technology and cybersecurity. The results of this restricted preview were significant enough to justify a broader public release — with important limitations.
What Makes Claude Fable Different?
Claude Fable is the public version of Claude Mythos — but it’s not the same model that Project Glasswing partners had access to.
The public version is expected to include substantial guardrails and have more limited cybersecurity functions than the version available to Project Glasswing partners.
In practical terms this means:
What Claude Fable can do:
- Advanced reasoning and complex multi-step tasks significantly beyond previous Claude models
- Significantly improved ability to handle complex multi-step tasks over long periods
- Defensive cybersecurity assistance — helping organisations identify and patch vulnerabilities
- Advanced coding and software development
- Complex document analysis and research
What Claude Fable won’t do: The guardrails specifically limit the offensive cybersecurity capabilities that made Claude Mythos controversial in its restricted form. The model that could autonomously discover and chain zero-day exploits across major operating systems is not the model being released to the public.
How Much Does Claude Fable Cost?
Access to Claude Fable is priced at roughly double the current Claude Opus tiers. The original Mythos pricing shared in April was five times higher than Opus. The public version is cheaper but still double anything available today.
This positions Claude Fable as a premium professional and enterprise product — not a consumer upgrade to Claude Pro. It’s targeting organisations and professional users who need significantly more capability than current models offer and are willing to pay a premium for it.
Why Was Claude Mythos Considered Too Dangerous to Release?
The cybersecurity capabilities of Claude Mythos in its full form are genuinely significant. The model demonstrated the ability to autonomously discover and chain zero-day exploits — previously unknown security vulnerabilities — across major operating systems and browsers.
This dual use problem is at the heart of why Anthropic delayed public release. The same capability that makes the model extraordinarily valuable for defensive security — finding vulnerabilities before malicious actors do — could also be misused to attack systems if available without restriction.
The solution Anthropic settled on — releasing a capability limited public version while maintaining the full model for vetted Project Glasswing partners — is a compromise that attempts to balance utility and safety.
What Does This Mean for the AI Landscape?
Claude Fable’s release marks a significant moment in AI development. It represents the first time a model specifically designed around advanced cybersecurity capabilities has been made broadly available — even in a limited form.
For comparison — ChatGPT and Google Gemini are general purpose AI assistants. Microsoft Copilot is focused on productivity. Claude Fable occupies a different category — an AI model built with professional and enterprise capability at its core, priced accordingly.
The release also raises the competitive bar significantly. If Claude Fable delivers on its promise of substantially improved complex reasoning and multi-step task handling — other AI companies will need to respond.
Should You Upgrade to Claude Fable?
For most individual users — probably not immediately. At double the price of current Claude Opus models Claude Fable is positioned for professional and enterprise users with specific high complexity needs.
For organisations in cybersecurity, software development, legal, financial services or research — the significantly improved capability for complex multi-step tasks may well justify the premium.
For everyday users currently on Claude Pro — the existing models remain excellent for writing, analysis and everyday tasks. Claude Fable is for users who are consistently hitting the limits of what current models can do.
Anthropic’s Broader Momentum
Claude Fable’s release comes at a remarkable moment for Anthropic. The company recently closed a $65 billion funding round at a $965 billion valuation — and is reportedly preparing for an IPO later in 2026. Claude Fable is the product statement that accompanies that financial momentum — a demonstration that Anthropic’s research investment is producing capabilities that justify its extraordinary valuation.
Recommended Products on Amazon
For professionals and developers exploring Claude Fable:
- Monitor for Home Office — more screen space makes working with advanced AI models across complex tasks significantly more efficient
- Blue Yeti Microphone — essential for professionals using AI tools for meetings, presentations and technical work
- The Coming Wave by Mustafa Suleyman — essential reading for understanding the capabilities and implications of models like Claude Fable
This article contains affiliate links. If you purchase through our links we may earn a small commission at no extra cost to you. This helps us keep this website free and updated.