Claude Fable 5 announcement hero image
AI
MGB Max craft  

Anthropic Just Released Its Most Powerful AI to the Public — Meet Claude Fable 5

June 10, 2026 · 5 min read · #Anthropic #AI #Claude


Two months ago, Anthropic unveiled Mythos — a model so capable at finding and exploiting software vulnerabilities that the company refused to release it publicly. Instead, it was quietly handed to a handful of vetted cybersecurity firms through a restricted program called Project Glasswing.

Today, that changes. Sort of.

Anthropic just launched Claude Fable 5: a Mythos-class model, with guardrails, available to everyone. And it might be the most significant AI release of the year.


What Is Claude Fable 5?

Fable 5 is best understood as Mythos with a safety layer on top. Same underlying model, same frontier capabilities — but with classifiers that intercept queries in high-risk areas like cybersecurity, biology, and chemistry, routing them to Claude Opus 4.8 instead.

Anthropic describes it as “a Mythos-class model that we’ve made safe for general use,” with capabilities that “exceed those of any model we’ve ever made generally available.”

The result is a model that’s state-of-the-art on nearly every benchmark — with a few asterisks where the safety fallbacks kick in.


The Benchmarks: A Clear Step Forward

Claude Fable 5 benchmark comparison table
Claude Fable 5 benchmark results vs. competitors. © Anthropic

The numbers are striking, especially on coding:

  • SWE-Bench Pro (real-world software engineering): Fable 5 scores 80.3% vs 69.2% for Opus 4.8, 58.6% for GPT-5.5, and 54.2% for Gemini 3.1 Pro
  • FrontierCode Diamond (harder coding set): 29.3% vs 13.4% for Opus 4.8 — more than double
  • Strong improvements across knowledge work, spatial reasoning, tool use, legal, and health benchmarks

The practical implication: Anthropic reported that a Ruby codebase with 50 million lines was migrated in a day — a task that would have taken an entire team more than two months to complete by hand.


Two Models, One Release

Alongside Fable 5, Anthropic is also shipping Claude Mythos 5 — the same model with some safeguards removed. Claude Mythos 5 is restricted to vetted partners working on cyberdefense and infrastructure, and is not generally available. It’s an upgrade to Mythos Preview through Project Glasswing.

Anthropic says Fable 5 has the strongest cybersecurity capabilities of any publicly available model, while Mythos 5 claims the strongest cybersecurity capabilities of any model, period.


How to Access Fable 5

Fable 5 is available immediately on the Claude API (model ID claude-fable-5), Claude Platform, Claude Code, consumption-based Enterprise plans, and on AWS, Google Cloud, and Microsoft Foundry.

Pricing: $10 per million input tokens, $50 per million output tokens — double Opus 4.8, with a 90% discount via prompt caching.

For subscription users, the rollout is more nuanced. Through June 22, Fable 5 is included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost. On June 23, Anthropic will remove it from those plans, requiring usage credits — with plans to restore it as a standard subscription feature as soon as possible.


The Bigger Picture: A Company Walking a Tightrope

Fable 5’s launch comes with a striking amount of self-awareness from Anthropic.

The release follows Anthropic’s public plea urging major global AI labs to establish a coordinated brake pedal on frontier AI development — warning that systems are advancing so rapidly they may soon achieve recursive self-improvement, autonomously improving themselves without human intervention.

In other words: Anthropic is simultaneously releasing its most powerful public model ever, and warning the world that models like it might be getting too powerful to control. It’s a tension the company hasn’t fully resolved — and probably can’t.

The timing also matters commercially. Fable 5’s launch comes as Anthropic prepares to enter the public markets, alongside OpenAI and Elon Musk’s SpaceX — capitalizing on growing investor interest ahead of a potentially massive IPO expected as soon as this year.


The Bottom Line

Claude Fable 5 is the most capable publicly available AI model as of today. The coding benchmarks alone are remarkable — and if the real-world performance matches the numbers, it’s a genuine leap forward for developers.

The catch: the safety guardrails mean it’s not the full Mythos experience. For most users, that’s fine. For those who need frontier cybersecurity capabilities, Project Glasswing is still the only path.

Fable 5 is available right now. Mythos 5 isn’t — unless Anthropic already trusts you with it.


Sources: Anthropic, VentureBeat, TechCrunch, CNBC, SD Times

Leave A Comment