Anthropic Launches Claude Fable 5: The First Public Mythos-Class Model

Anthropic announced on Tuesday, June 9, 2026, the launch of its new model Claude Fable 5, a step that marks the first public availability of a Mythos-class model — the tier that sits above the company's Opus class in capability. Anthropic confirmed that Fable 5's capabilities exceed any model it has previously made generally available, and that it is state-of-the-art on nearly all tested capability benchmarks.

Two Models in One Launch

Alongside Fable 5, Anthropic unveiled a second model called Claude Mythos 5 — the same underlying model but with some safeguards lifted. Mythos 5 remains available only to a limited group of cyber defenders and critical infrastructure providers within "Project Glasswing," in collaboration with the US government, with access to be expanded later through a trusted access program. The core difference between the two models is the safety layer: Fable comes with safeguards for general use, while Mythos is more open for approved organizations.

A Leap in Performance Across Domains

According to the company's announcement, Fable 5 shows exceptional performance in software engineering, knowledge work, computer vision, and scientific research. In early testing, Stripe reported that the model compressed months of engineering work into days, performing a codebase-wide migration in a 50-million-line Ruby codebase in a single day — work that would have taken a full team over two months by hand. The model is also more token-efficient than previous models.

On the vision front, Fable 5 became the new state-of-the-art for visual tasks, able to extract precise numbers from scientific figures and rebuild a web app's source code from screenshots alone. The more powerful Mythos 5 version recorded breakthroughs in areas such as drug design and molecular biology, described by the company as its first model to consistently produce novel, compelling scientific hypotheses.

Strict Safety Safeguards

Given the advanced capabilities of this class of models, especially in cybersecurity, Anthropic launched the model with a layer of safety classifiers that detect misuse attempts in sensitive areas such as cybersecurity, biology, and chemistry. When such a request is detected, the response automatically falls back to the less-capable Claude Opus 4.8 model rather than an outright refusal, with the user informed. The company noted that more than 95% of sessions require no fallback at all.

Anthropic said it conducted extensive testing for resistance to jailbreak attempts, including an external bug bounty that ran for over a thousand hours without finding a universal jailbreak.

Pricing and Availability

Claude Fable 5 is available today via the Claude API under the model identifier claude-fable-5, priced at 10 dollars per million input tokens and 50 dollars per million output tokens — roughly double the price of Opus 4.8. On subscription plans, the model is included at no extra cost on Pro, Max, Team, and Enterprise plans through June 22, after which usage will require usage credits, with the company working to restore it as a standard part of subscriptions as capacity allows.

What Does This Mean for Developers?

Fable 5 represents a strong option for long, complex tasks that were out of reach for previous models, especially in agentic coding, building complete applications, and document analysis. That said, the high cost and conservative safeguards remain practical considerations when choosing it for a production workflow, as the safeguards may sometimes route legitimate requests to the fallback model.