Anthropic: Claude Fable 5.1 (batch)

Claude Fable 5.1 (batch) is a text model from Anthropic, offered on api14 under the model id anthropic/claude-fable-5.1:batch. It is the successor to Claude Fable 5, and Anthropic describes it as an improvement across the board, with the largest gains in agentic coding, long-running agentic workflows, and knowledge work. The batch label marks this as the batch variant of the model, which is the endpoint you call when you want to route work through api14's batch handling rather than an interactive request.

The model accepts text, images, and files as input, and it produces text as output. Its context window is 1,000,000 tokens, and the maximum output length is 128,000 tokens. That combination of a very large input window and a large output budget means you can hand it long documents, large codebases, or extended conversation histories in a single request and still get a substantial response back. Supported request parameters include max_tokens, stop, response_format, structured_outputs, tool_choice, tools, reasoning, reasoning_effort, include_reasoning, and verbosity, so you can control output shape, tool use, and how much reasoning detail is returned.

Those capabilities point to a few practical workloads. Long code refactors fit naturally, since the model can take in a large portion of a repository and return rewritten code within the output limit. Front-end and visual work is supported by image input alongside text. Agentic workflows that run for many steps benefit from the large context, because intermediate state and tool results can accumulate without forcing you to truncate. Knowledge work such as summarising, extracting, or transforming long documents is also a direct fit, and structured_outputs plus response_format let you constrain the result to a schema your application can parse.

Calling it through api14 works the same way as any other OpenAI-compatible model: send a request to the chat completions endpoint with the model field set to anthropic/claude-fable-5.1:batch, and pass your messages in the standard format. Because the API is OpenAI-compatible, existing clients and SDKs that target OpenAI can be pointed at api14 with minimal changes. You can attach tools and tool_choice for function calling, set reasoning and reasoning_effort to tune how the model thinks through a problem, use include_reasoning if you want the reasoning surfaced, and adjust verbosity to control how expansive the prose output is. max_tokens caps the response, stop sets stopping sequences, and structured_outputs or response_format enforce a particular output structure.

One thing to keep in mind is that this is a batch variant, so it suits workloads where you submit work and collect results rather than needing a tightly interactive loop. If your task is small, short-lived, or needs immediate turn-by-turn responses, a non-batch model may be a better fit. The 128,000 token output ceiling is generous but still finite, so very large single outputs may need to be split across calls. Within those bounds, the model is aimed at developers doing serious coding, agentic, and document-heavy work.