What Claude Fable 5 and Mythos-Class AI Mean
Claude Fable 5 is a new Mythos-class AI model from Anthropic that offers advanced reasoning, coding, and vision capabilities while adding safety routing so it can be used by mainstream users for everyday work. It is now the most capable Claude model available to the public, and it is designed to handle complex tasks across software engineering, knowledge work, scientific research, and long-running workflows. Mythos-class AI refers to Anthropic’s highest capability tier, which until now was reserved for selected cyberdefenders and infrastructure providers under Project Glasswing. With Fable 5, Anthropic takes the same underlying model used in Claude Mythos 5 and adds guardrails so general users can access top-end performance without direct exposure to the riskiest capability areas that remain gated behind trusted-access programs.
How Fable 5 Differs from Previous Claude Models
Claude Fable 5 builds on Anthropic AI models like Claude Opus 4.8 by improving performance on benchmarks tied to real work, including code migration, financial reasoning, and scientific vision tasks. Anthropic reports that Fable 5 outperformed Opus 4.8 as well as competing large models in its internal testing, with examples such as extracting precise values from scientific figures and rebuilding source code from screenshots. One quotable example shows how this translates into outcomes: according to Anthropic, Stripe used Fable 5 to complete a migration across a 50‑million‑line Ruby codebase in around one day, compared with more than two months of manual engineering work. The model’s strengths also include long-context reasoning and persistent agent workflows, especially when used through tools like Claude Code and managed platforms such as Amazon Bedrock and Claude Platform on AWS.
Safety Routing: Why It Matters for Everyday Use
Safety routing is the core feature that makes a Mythos-class AI like Claude Fable 5 suitable for everyday users instead of only specialist teams. Fable 5 shares its base model with Claude Mythos 5, but when it detects prompts related to cybersecurity, biology, chemistry, or model distillation, it routes those requests to Claude Opus 4.8 instead. Users are informed when this fallback happens, and Anthropic says it occurs in less than 5% of sessions on average, so most interactions still run on Fable 5. This design keeps high-capability features available for typical work while reducing exposure to sensitive, higher-risk outputs. Mythos 5, by contrast, relaxes some of these safeguards and is reserved for approved partners through Project Glasswing and future trusted-access programs focused on cybersecurity and biology research.
Access, Pricing Shift, and Data Retention Trade-offs
Claude Fable 5 is available to paid Claude Pro, Max, Team, and seat-based Enterprise subscribers at no additional cost through June 22, turning launch day into a discounted access window before usage-credit pricing begins. After this period, both Fable 5 and Mythos 5 move to USD 10 (approx. RM46) per million input tokens and USD 50 (approx. RM230) per million output tokens, which makes Mythos-class AI an explicit cost–benefit decision for teams weighing larger context and agent workflows against price. Anthropic is also introducing a 30-day data retention policy for Fable 5, Mythos 5, and similarly capable future models. The company says this traffic will not be used to train new Claude models or for non-safety purposes, but security-sensitive enterprises must consider how that retention rule interacts with their legal, compliance, and procurement requirements.
A New Phase of Accessible High-End AI
Fable 5 marks a shift in how Anthropic exposes its top-tier Anthropic AI models to the wider public. Mythos-level capabilities that were previously limited to a small group of cyberdefenders and infrastructure providers are now available inside a product meant for day-to-day work, with safety routing acting as a buffer around risky domains. The combination of strong performance in software engineering, knowledge work, and scientific tasks with limited-time discounted access on June 22 positions Claude Fable 5 as a flagship option for users who want Mythos-class AI without joining a restricted program. At the same time, data retention, fallback routing, and a defined price per million tokens signal that high-end AI is becoming a managed service with clear trade-offs, not an unrestricted experiment running in the background.
