Anthropic has launched Claude Fable 5, the first Mythos-class model featuring a 1-million-token context window and an 80.3% score on SWE-bench Pro.
Anthropic has announced the release of Claude Fable 5, its first generally available "Mythos-class" large language model designed for complex, long-horizon agentic tasks. Featuring a 1-million-token context window and supporting up to 128,000 output tokens, the model has set a new state-of-the-art benchmark with an 80.3% score on SWE-bench Pro. Priced at $10 per million input tokens and $50 per million output tokens, Claude Fable 5 incorporates strict safety safeguards that route high-risk queries in domains like cybersecurity and biology to older models, making this powerful frontier-tier AI accessible for general developers.
Anthropic's Claude Fable 5 signals that the frontier model race is shifting from raw parameter size to long-horizon agentic capabilities and safety-gated routing.
- –**Agentic Powerhouse:** An 80.3% score on SWE-bench Pro indicates that Fable 5 is highly optimized for autonomous software engineering, a key differentiator for development workflows.
- –**Massive Context & Output:** A 1M input context coupled with 128k output tokens enables extremely long-running agentic loops without context loss, though users must manage potentially high token consumption costs.
- –**Tiered Safety Routing:** By routing sensitive queries to Claude Opus 4.8, Anthropic is pioneering a hybrid deployment strategy to balance raw capability with strict compliance and risk mitigation.
DISCOVERED
1h ago
2026-06-12
PUBLISHED
1h ago
2026-06-12
RELEVANCE
AUTHOR
slashdevhq