Hajra Naz
Anthropic has launched Claude Fable 5, a Mythos-class AI model designed for general use with added safety layers. The company says it is the most capable Claude model released for public access, with strong results across software engineering, knowledge work, vision tasks, scientific research, memory, and long-context reasoning.
Safety Guardrails
Anthropic says Fable 5 runs on the same base model as Claude Mythos 5 but includes stricter safety controls for public use.
When the system detects sensitive requests related to cybersecurity, chemistry, biology, or model distillation, it routes the response to Claude Opus 4.8 instead. Users receive a notice when this happens.
The company notes that these safeguards may block some safe prompts, though fallback usage stays below 5 percent on average.
Claude Mythos 5
Anthropic is also releasing Claude Mythos 5 for a limited group of approved users.
Mythos 5 uses the same base model as Fable 5 but reduces certain restrictions in controlled settings. The first rollout goes to Project Glasswing partners working with the US government.
The company plans wider access later through a trusted program for cybersecurity teams and a separate channel for selected biology researchers.
Read More: Anthropic’s Model Context Protocol has crossed 97 million installs in just 16 months
Performance Claims
Anthropic says Fable 5 outperforms earlier Claude models across multiple benchmarks.
In software engineering, Stripe reportedly used Fable 5 to complete a full migration across a 50-million-line Ruby codebase in one day. The same task would normally take a large engineering team over two months.
The model also shows strong results in coding, finance reasoning, chart interpretation, scientific vision tasks, and long-context analysis.
Research Uses
Anthropic reports strong early results for Mythos 5 in life sciences.
Internal teams used it to speed up parts of protein design workflows, cutting certain drug discovery steps by about ten times.
In blinded evaluations, Anthropic scientists preferred Mythos 5 outputs over Opus-class responses in most cases involving molecular biology tasks.
Data Retention
Anthropic is introducing a 30-day data retention policy for Fable 5, Mythos 5, and future high-capability models.
The company says stored data will support safety monitoring, including jailbreak detection and misuse analysis. It will not be used for training future Claude models.
Read More: Anthropic Introduces Code Review Tool to to Check Rising AI-Generated Code
Pricing and Availability
Claude Fable 5 is available now through the Claude API under the name claude-fable-5.
Claude Mythos 5 remains limited to Glasswing partners for now. Both models are priced at $10 per million input tokens and $50 per million output tokens.
Fable 5 is included in selected subscription plans until June 22. After that date, access will shift to usage credits, with possible extensions depending on capacity.
