Claude Fable 5 lands on Vercel AI Gateway

Claude Fable 5 from Anthropic is now available on Vercel's AI Gateway. Classified as a Mythos-class model, Fable 5 is described as a notable step up over prior Claude models specifically on long-running, ambiguous, and multi-step tasks — executing end-to-end on work that previously required frequent human check-ins. The model sustains productive output across multi-day runs, dependably dispatches parallel sub-agents, and lower effort settings are said to often match what prior Claude models produced at their highest effort. Code review, bug-finding, and repository investigation capabilities are noted as stronger, and first-shot correctness on complex problems is described as noticeably higher.

Fable 5 ships with blocking classifiers that refuse offensive cybersecurity, biology, and summarized-thinking extraction requests, with Anthropic citing real misuse risk in those areas.

Fable 5 ships with blocking classifiers that refuse offensive cybersecurity, biology, and summarized-thinking extraction requests, with Anthropic citing real misuse risk in those areas. Anthropic also does not support Zero Data Retention for this model, as some misuse patterns only become visible across cumulative requests; prompts and completions are retained for 30 days and are not used to train Claude. Developers can access the model by setting the model identifier to `anthropic/claude-fable-5` in the AI SDK.

Vercel's AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations. It includes custom reporting, Zero Data Retention support (for other models), and dynamic provider sorting by latency and cost. The gateway reflects provider pricing with no markup and charges no platform fee on inference, including on Bring Your Own Key (BYOK) requests.

Key facts

1. 01Claude Fable 5 is a Mythos-class model from Anthropic, now available on Vercel AI Gateway.
2. 02It is described as a notable step up over prior Claude models on long-running, ambiguous, multi-step tasks.
3. 03The model sustains output across multi-day runs and dependably dispatches parallel sub-agents.
4. 04Lower effort settings on Fable 5 are said to often match prior Claude models at their highest effort.
5. 05Blocking classifiers refuse offensive cybersecurity, biology, and summarized-thinking extraction requests.
6. 06Anthropic does not support Zero Data Retention for Fable 5; prompts and completions are retained for 30 days but not used for training.
7. 07To use the model, developers set the model identifier to `anthropic/claude-fable-5` in the AI SDK.

Topics