Anthropic launches Claude Fable 5
Anthropic has launched Claude Fable 5, a new "Mythos-class" model designed for complex agentic workflows, software engineering, and research synthesis. The model is available via the Claude API, subscription plans, and cloud platforms, with safety guardrails that fallback to Claude Opus for risky queries.
Claude Fable 5 represents a clear pivot towards autonomous agentic AI, but its restricted sister model Mythos 5 highlights a growing industry trend where the absolute most powerful, unrestricted capabilities are gated for vetted partners while the general public gets a safer, rate-limited edition.
* Fable 5 shares the underlying architecture of the restricted Mythos 5 but implements automatic fallbacks to Claude Opus if safety guardrails are triggered.
* Strong focus on software engineering and long-horizon tasks, achieving 80.3% on SWE-Bench Pro.
* Premium pricing at $10/M input and $50/M output indicates that frontier reasoning models remain expensive to run.
DISCOVERED
1h ago
2026-06-10
PUBLISHED
1h ago
2026-06-10
RELEVANCE
AUTHOR
FutureStacked