Anthropic Debuts Next-Gen Mythos-Class Models: Claude Fable 5 and Mythos 5 Set New Industry Standards for AI Autonomy

Anthropic has officially launched its next-generation Mythos-class models, Claude Fable 5 and Claude Mythos 5, which deliver state-of-the-art performance in coding, vision, and scientific research. While Fable 5 is engineered for safe general use, Mythos 5 provides specialized, high-capacity capabilities for critical cybersecurity and biomedical infrastructure through a trusted access program.

RMN Digital AI Desk
New Delhi | June 10, 2026

A New Benchmark in AI Capability

On June 9, 2026, Anthropic announced the release of Claude Fable 5 and Claude Mythos 5, marking the arrival of the “Mythos-class” tier of models that sit above the previous Opus-class in capability. These models represent a significant leap in AI autonomy, demonstrating the ability to work autonomously for longer periods than any previous Claude iterations.

Fable 5 has achieved state-of-the-art status across nearly all tested benchmarks. Its performance in software engineering is particularly notable; early testing by Stripe showed that the model could perform a codebase-wide migration on a 50-million-line Ruby codebase in a single day—a task that would typically take a human team over two months. Furthermore, Fable 5 has demonstrated unprecedented vision capabilities, such as rebuilding web app source code from screenshots and beating the game Pokémon FireRed using only raw game screenshots without additional tools or game-state information.

Fable 5 compressed months of engineering into days, performing a codebase migration on 50 million lines of code in just 24 hours.

Dual-Track Innovation: Fable vs. Mythos

The two models share the same underlying architecture but differ in their safety configurations.

• Claude Fable 5: This model is designed for general use and is equipped with a robust set of safety classifiers. These classifiers monitor for potential misuse in areas like cybersecurity, biology, and chemistry; if a query is deemed high-risk, the system automatically falls back to the Claude Opus 4.8 model to ensure safety while maintaining utility.
• Claude Mythos 5: Initially deployed through Project Glasswing in collaboration with the U.S. government, Mythos 5 is reserved for a select group of cyberdefenders and researchers. It features the strongest cybersecurity capabilities of any model globally, with specific safeguards lifted to allow for high-level defensive work and complex drug design research.

Scientific and Financial Breakthroughs

Beyond coding, the Mythos-class models are transforming life sciences and finance. Mythos 5 has been used by internal experts to accelerate drug design by tenfold, matching or beating skilled human operators in selecting binding sites and running protein design tools. In molecular biology, the model is the first to consistently produce novel scientific hypotheses, several of which have already been advanced to experimental evaluation.

At less than half the price of previous versions, Anthropic’s Mythos-class models are democratizing state-of-the-art AI while maintaining rigorous safety standards.

In the financial sector, Fable 5 scored the highest of any model on Hebbia’s Finance Benchmark for senior-level reasoning. It excels at document-based reasoning, chart interpretation, and root-cause analysis, making it a powerful tool for complex analytical tasks that require high attention to nuance.

Industry Impact and Accessibility

Anthropic is positioning these models as both more powerful and more affordable. Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens, which is less than half the price of the preceding Claude Mythos Preview.

Starting today, Fable 5 is available on the Claude API and for consumption-based Enterprise plans. For subscription-based plans (Pro, Max, Team), Fable 5 is included at no extra cost through June 22, after which it will transition to a credit-based system until capacity allows for permanent inclusion.