Anthropic Releases Claude Fable 5 and Restricted Mythos 5 Model

June 9, 2026
Anthropic Releases Claude Fable 5 and Restricted Mythos 5 Model

Anthropic today launched Claude Fable 5, a new Mythos-class AI model designed for general availability, alongside Claude Mythos 5, a version with fewer safeguards that will initially be limited to cybersecurity and infrastructure partners through Project Glasswing.

The company says Fable 5 is its most capable publicly available model to date, delivering improvements across software engineering, knowledge work, vision, scientific research, memory, and long-running autonomous tasks. Because of the model’s advanced capabilities, Anthropic built new safeguards that automatically route certain cybersecurity, biology, chemistry, and distillation-related requests to Claude Opus 4.8 instead.

Anthropic said those safeguards were intentionally configured conservatively to support a faster public release. According to the company, fallback responses occur in fewer than 5% of sessions on average, while more than 95% of Fable 5 interactions receive responses directly from the new model.

For organizations requiring higher-risk capabilities, Anthropic is also releasing Claude Mythos 5. The model shares the same underlying architecture as Fable 5 but removes some of the restrictions around cybersecurity-related tasks. Access is currently limited to participants in Project Glasswing, a program developed in collaboration with the U.S. government to support cyber defenders and critical infrastructure providers.

The company highlighted several benchmark and real-world performance gains. In software engineering, Anthropic said Stripe used Fable 5 to complete a migration across a 50-million-line Ruby codebase in a single day, a task that would otherwise have required a team working for more than two months. The model also achieved the highest score among frontier models on Cognition’s FrontierCode evaluation.

Anthropic reported strong results in analytical work as well. Fable 5 achieved the highest score on Hebbia’s Finance Benchmark for senior-level reasoning, while trading firm IMC said the model performed strongly across factual lookup, conceptual reasoning, root-cause analysis, and expected-value analysis tasks.

Vision capabilities were another focus of the release. Anthropic said Fable 5 can extract precise information from scientific figures, recreate web application source code from screenshots, and complete complex visual tasks with less supporting infrastructure than previous Claude models. The company noted that Fable 5 completed Pokémon FireRed using only raw game screenshots and a vision-based setup.

The model also introduces improvements in long-context reasoning and memory. Anthropic said Fable 5 remains effective across tasks involving millions of tokens and benefits significantly from persistent notes and file-based memory during extended workflows.

Much of the announcement focused on scientific research. Anthropic said internal teams used Mythos 5 to accelerate parts of the drug discovery process by roughly tenfold. In one study, the model independently completed protein design tasks typically handled by researchers, including selecting binding sites, operating design tools, and recovering from failures during the process.

The company also said Mythos 5 generated molecular biology hypotheses that scientists preferred over Opus-class model outputs roughly 80% of the time in blinded comparisons. Anthropic noted that one hypothesis involving an E. coli protein was later corroborated by an independent research group.

Anthropic said the safeguards underwent extensive testing before launch, including external red-team exercises and a bug bounty program. According to the company, researchers did not discover any universal jailbreaks during more than 1,000 hours of testing.

Anthropic is also implementing a new data retention policy for Mythos-class models and future systems with similar capability levels. Business customer traffic will be retained for 30 days for safety-related purposes, though the company said the data will not be used to train future Claude models and will be deleted after the retention period in most cases.

Claude Fable 5 is available today through the Claude API and Anthropic’s platforms. Both Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens. Anthropic said access for subscription users will be rolled out in stages as it manages demand and available capacity.

See more here:

This analysis is based on reporting from anthropic.

Image courtesy of Anthropic.

This article was generated with AI assistance and reviewed for accuracy and quality.

Last updated: June 9, 2026

About this article: This article was generated with AI assistance and reviewed by our editorial team to ensure it follows our editorial standards for accuracy and independence. We maintain strict fact-checking protocols and cite all sources.

Word count: 655Reading time: 0 minutes

AI Tools for this Article

Browse All Articles
Share this article:
Next Article

AI News Daily

Breaking Intelligence • Since 2023

Join hundreds of thousands of AI professionals who start their day with our curated newsletter. Get breaking news, expert analysis, and exclusive insights.

Stay Ahead of AI

Get the latest AI breakthroughs, tools, and insights delivered to your inbox every week.

Free forever Unsubscribe anytime No spam guarantee

Go Premium

Unlock unlimited AI tools and an ad-free reading experience designed for AI professionals.

• Ad-free experience• Premium AI tools
Start Free Trial

14-day free trial • Cancel anytime
Plus $9/mo • Pro $90/yr (2 months free)

Follow Our Community

ChatAI

Breaking Intelligence

Your daily briefing on what matters in AI. Trusted by developers, researchers, executives, and AI enthusiasts worldwide.

© 2026 ChatAI. All rights reserved.