Claude Fable 5 — Anthropic's Most Powerful Model Is Now Public

Anthropic just released its most powerful AI model to the public — and it comes with a built-in safety system unlike anything we have seen before.

Claude Fable 5 launched on June 9, 2026, as the first Mythos-class model available to everyday users. It scores more than 10% higher than Claude Opus 4.8 across key benchmarks in software engineering, scientific research, and autonomous task execution. It can also work independently for longer stretches than any previous Claude model.

The Safety Split

What makes this release unusual is how Anthropic handled the risk. Rather than releasing one model, they released two.

Fable 5 is the public version. It includes safety classifiers — separate AI systems that monitor conversations in real time. When a query touches sensitive areas like cybersecurity exploits, dangerous chemistry, or biological risks, the response is quietly handed off to the less capable Opus 4.8 model. Anthropic says this fallback triggers in fewer than 5% of sessions.

Mythos 5 is the unrestricted version of the same model. It is only available to vetted partners through Project Glasswing, Anthropic’s programme for cybersecurity researchers and scientific institutions.

This approach — same model, different guardrails — is a new pattern in AI safety. Instead of refusing a request entirely, the system downgrades its response.

Pricing and Access

Fable 5 is free on Claude Pro, Max, Team, and Enterprise plans until June 22. After that, it moves to usage-based pricing at $10 per million input tokens and $50 per million output tokens through the API.

If you have been using Claude Opus 4.8, you will notice a significant jump in capability — particularly for coding, research, and complex reasoning tasks.

What This Means for You

If you use Claude regularly, try Fable 5 now while it is still included in your subscription. It is the strongest model Anthropic has ever offered publicly, and the free window closes on June 22.

For developers, the API pricing puts it at the premium end but below what many expected for a frontier model. Check our complete guide to Claude to get the most out of it, or see how it compares in our ChatGPT vs Claude vs Gemini breakdown. New to AI? Start with our beginner’s guide.