Anthropic releases Claude Fable 5, a Mythos-class model, now available on AI Gateway.
- •Excels at long-running, ambiguous, multi-step tasks with end-to-end execution without frequent human check-ins
- •Sustains productive output across multi-day runs and reliably dispatches parallel sub-agents
- •Stronger at code review, bug-finding, and repository investigation with noticeably higher first-shot correctness
- •Includes safety classifiers blocking offensive cybersecurity and biology content, with 30-day data retention policy
- •Accessible via AI Gateway unified API using model ID anthropic/claude-fable-5 with adaptive thinking support
This summary was automatically generated by AI based on the original article and may not be fully accurate.