Anthropic Launches Claude Fable 5 With Unprecedented Coding and Vision Capabilities

Anthropic has launched Claude Fable 5 and Claude Mythos 5, bringing major advances in software engineering, vision, and knowledge work to both public and enterprise users. The release introduces Mythos, a new class of Claude models that Anthropic says surpasses its Opus family in capability, while implementing new safeguards designed to prevent misuse.



The launch holds particular significance for developers and security teams within the Apple ecosystem. Apple recently integrated Anthropic's models into Xcode 26.3 to power autonomous coding agents. Additionally, Apple participates in Project Glasswing, Anthropic's initiative to help cyber defenders identify and address vulnerabilities in critical software infrastructure.


Fable 5 is designed for general use and delivers state-of-the-art performance across most major AI benchmarks. In early testing, companies like Stripe reported that Fable 5 compressed months of software engineering work into a matter of days, successfully migrating a massive codebase autonomously. The model also delivered leading scores on senior-level financial reasoning and data analysis benchmarks.

It excels at complex, long-running tasks and demonstrates remarkable vision capabilities. According to Anthropic, the model can extract precise numbers from scientific figures and completely rebuild a web application's source code using nothing but screenshots. It also completed Pokémon FireRed using only raw game screenshots, without maps, navigation aids, or game-state information.

Because the underlying technology is so advanced, Fable 5 includes aggressive safety classifiers. If classifiers detect requests related to cybersecurity, biology and chemistry, or model distillation, the query is automatically handled by Claude Opus 4.8 instead. Anthropic notes these safeguards are currently tuned conservatively and trigger in less than five percent of sessions on average.

For verified cyber defenders and infrastructure providers, Anthropic released Claude Mythos 5. This version shares the same underlying model as Fable 5 but has safeguards lifted in certain areas, particularly cybersecurity. It is initially deploying through Project Glasswing to assist government and enterprise security teams in securing critical software, with plans for a broader trusted access program in the future.


Beyond cybersecurity, Anthropic highlighted Mythos 5's capabilities in scientific research. The company says the model can generate novel molecular biology hypotheses, assist with protein design, and conduct autonomous genomics research. In internal testing, Anthropic found that Mythos 5 accelerated aspects of the drug design process by roughly tenfold and produced biology hypotheses that researchers preferred over Opus-class models approximately 80 percent of the time.

Both Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens. Developers can access Fable 5 through the Claude API today. Subscription users on Pro, Max, Team, and seat-based Enterprise plans will have access at no extra cost through June 22, after which it will transition to a usage credit system until Anthropic expands available capacity.