Anthropic has launched Claude Fable 5, the first Mythos-class model made broadly available to the public.
Anthropic has publicly released Claude Fable 5, representing its first Mythos-class AI model made generally available. Positioned above the Claude Opus series, the new model is designed for highly complex, long-running agentic tasks such as multi-step planning, tool utilization, and self-verification of outputs, scoring 80.3% on SWE-Bench Pro. To maintain safety standards while providing this advanced tier of intelligence, Anthropic implemented a fallback guardrail system that automatically reroutes sensitive queries involving cybersecurity, biology, or chemistry to Claude Opus 4.8.
Anthropic's release of Claude Fable 5 signals a massive push into autonomous agentic workflows, though the strict fallback mechanism highlights the ongoing challenges of deploying next-generation models safely.
- –The 80.3% score on SWE-Bench Pro indicates a major leap in coding capability, making Fable 5 a powerful tool for complex software engineering tasks.
- –The fallback redirect to Opus 4.8 for sensitive domains is an innovative compromise, allowing Anthropic to ship a highly capable model to the public while mitigating catastrophic risks.
- –At $10/M input and $50/M output tokens, the model is priced as a premium tier, indicating that agentic workflows will remain relatively expensive for high-volume use cases.
DISCOVERED
2h ago
2026-06-11
PUBLISHED
2h ago
2026-06-11
RELEVANCE
AUTHOR
Humpidumpi2