Claude Fable 5 access restored on AI Gateway
By Steven Van ·
The model returns after US export controls were lifted, with tougher safety classifiers that may flag routine coding requests.
Anthropic's Claude Fable 5, the Mythos-class model, is back on Vercel's AI Gateway, after the US Government lifted the export controls that had suspended it. It's the same model that was briefly available between June 9 and June 12; what's changed is the safety classifiers, which are now updated and more robust.
Because of that, routine requests such as coding and debugging may occasionally trigger a refusal in the near term. Vercel recommends configuring model fallbacks so that if Anthropic declines a request to Fable 5, AI Gateway retries it against the next model in the list, such as Claude Opus 4.8 or Claude Sonnet 5. Fallbacks work across every API format. The model is called with the name anthropic/claude-fable-5.
Anthropic isn't offering Zero Data Retention for Fable 5, since some misuse patterns only show up across multiple requests rather than any single one. Prompts and completions are instead kept for 30 days and aren't used to train Claude.