AI BriefingClaude Fable 5 Now Available on AI Gateway
Key point
Anthropic's new model Claude Fable 5 has launched on Vercel AI Gateway, enhancing its ability to perform complex multi-step tasks.
Details
Anthropic's Mythos-class model, Claude Fable 5, has launched on AI Gateway. This model is optimized to perform longer, more ambiguous multi-step tasks than previous models, enabling end-to-end task execution without human intervention.
Key features include:
- Long-running capability: Maintains productive output even in tasks that run for days, and reliably controls parallel sub-agents.
- Improved performance: Enhanced code review, bug detection, and repository investigation capabilities, with a noticeably higher first-shot correctness rate on complex problems.
- Efficiency: Achieves performance comparable to the previous model's highest settings even with lower effort settings.
On the security and data policy side, blocking classifiers have been applied for cybersecurity, biology, and summarized-thinking extraction. Additionally, Zero Data Retention (ZDR) is not supported in order to detect abuse patterns through cumulative requests, and prompts and outputs are retained for 30 days but are not used for model training.
Users can start using it right away by setting the model name to anthropic/claude-fable-5 in the AI SDK.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.