Source: vercel.com/changelog/claude-fable-5-2-now-available-on-ai-gateway
Anthropic describes Fable 5.2 as built for open-ended technical work, with gains over earlier Fable models in coding, reasoning and vision. It can run long, ambiguous tasks unattended, and can act as an orchestrator that hands simpler work to smaller models or as an advisor that other Claude models consult. Effort ranges from low to max, with high as the API default.
Cybersecurity and biology safety classifiers are enabled on the model. Anthropic does not support Zero Data Retention for Fable 5.2, and prompts and completions are kept for 30 days but not used to train Claude.
To keep requests served when the cybersecurity classifier declines one, AI Gateway accepts a models array under providerOptions.gateway and tries each model in order. The model can be called through the AI SDK, the Chat Completions, Responses and Anthropic Messages APIs, or a coding agent connected to the gateway.
Why it matters if you build agents
Agent builders can call Fable 5.2 through AI Gateway and set a fallback chain so a classifier refusal does not end the run. Teams with Zero Data Retention enabled will get errors for this model and need a fallback route or a separate model. Check the 30-day retention policy before sending sensitive prompts.
Get stories like this every weekday
One short email with the day's agent tooling news.