The Claude Fable API gives you programmatic access to Anthropic's flagship Claude Fable models — the newer Claude Fable 5.1 (1M-token context, always-on adaptive thinking) and the earlier Claude Fable 5 — for complex reasoning, long-horizon agentic coding, and multimodal analysis, through one Muapi API key with both async and live-streaming endpoints, at Anthropic's own published token rates.
Submit a prompt (with an optional reference image and system prompt) and get back a request ID, then poll the shared result endpoint for the completed response.
Same model, delivered as server-sent events so tokens can be rendered to the user as they're generated instead of waiting for the full response.
Anthropic's earlier flagship model. Submit a prompt and get back a request ID, then poll the shared result endpoint for the completed response.
Same earlier-generation model, delivered as server-sent events for token-by-token rendering.
The Claude Fable API gives applications programmatic access to Anthropic's flagship Claude Fable models. Submit a prompt — optionally alongside a reference image and a system prompt to steer tone and behavior — and get back a text response. Claude Fable 5.1 is the current flagship: it runs with adaptive extended thinking always enabled, a 1 million token context window, and up to 128K tokens of output, purpose-built for long-horizon agentic coding, multistep research, and document/spreadsheet/slide work. Claude Fable 5 remains available for existing integrations that don't need the larger context window.
Muapi exposes both versions through an asynchronous REST endpoint for request/poll workflows and a Server-Sent Events streaming endpoint for token-by-token chat UIs. Compare Claude Fable against every other text model in the playground.
Claude Fable 5.1 defaults to a 1 million token context window with up to 128K output tokens, enough to hold an entire codebase or research thread in one call.
Extended thinking can't be turned off on either model, but its depth adapts automatically to the task — harder, multistep requests get more reasoning effort.
Pass an optional image_url alongside your prompt so Claude Fable can reason about charts, tables, PDFs, and screenshots, not just the text.
Set a system_prompt to steer tone, persona, or output format across a request.
Use the standard submit-and-poll endpoint for batch jobs and workflows, or the SSE streaming endpoint to render tokens live in a chat interface.
Call /claude-fable-5-1 for the newest flagship or /claude-fable-5 for the earlier release — same request shape, same Muapi API key.
Submit a JSON request with your Muapi API key, save the request ID, and poll the shared result endpoint — or call the /stream variant for a live SSE token feed. The request shape is identical for both model generations.
curl -X POST https://api.muapi.ai/api/v1/claude-fable-5-1 \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"Plan and outline a multi-file refactor for this module, then list the follow-up questions.","system_prompt":"You are a meticulous staff engineer.","image_url":"https://example.com/module-diagram.png"}'
# Response: {"request_id":"REQUEST_ID"}Call claude-fable-5-1 with a prompt, and optionally a system_prompt and image_url.
curl https://api.muapi.ai/api/v1/predictions/REQUEST_ID/result \ -H "x-api-key: YOUR_API_KEY" # Poll until status is completed, then read the generated text from outputs.
Poll until the task is completed, then read the generated text from outputs.
curl -X POST https://api.muapi.ai/api/v1/claude-fable-5-1/stream \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"Research and compare three approaches to this problem, citing tradeoffs."}'
# Response: a stream of SSE chunks, one per line:
# data: {"choices":[{"delta":{"content":"Here"}}]}
# data: {"choices":[{"delta":{"content":" are"}}]}
# data: [DONE]Call claude-fable-5-1/stream with the same fields to receive SSE chunks as the model generates them, ending in a [DONE] marker.
curl -X POST https://api.muapi.ai/api/v1/claude-fable-5 \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"Summarize the architectural tradeoffs in this diagram.","system_prompt":"You are a precise senior staff engineer.","image_url":"https://example.com/architecture.png"}'
# Response: {"request_id":"REQUEST_ID"}Call claude-fable-5 the same way for existing integrations that don't need the larger context window.
curl https://api.muapi.ai/api/v1/predictions/REQUEST_ID/result \ -H "x-api-key: YOUR_API_KEY" # Poll until status is completed, then read the generated text from outputs.
Same shared result endpoint, regardless of which model version submitted the task.
curl -X POST https://api.muapi.ai/api/v1/claude-fable-5/stream \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"Write a 600-word noir opening set in a server room."}'
# Response: a stream of SSE chunks, one per line:
# data: {"choices":[{"delta":{"content":"The"}}]}
# data: {"choices":[{"delta":{"content":" room"}}]}
# data: [DONE]Call claude-fable-5/stream for a live SSE token feed on the earlier model.
Claude Fable is Anthropic's flagship line of language models. Muapi exposes both the current release, Claude Fable 5.1, and the earlier Claude Fable 5 as standard REST APIs, so any application can call them with an API key instead of an Anthropic account.
Claude Fable 5.1 adds a larger 1M-token context window, always-on adaptive extended thinking steerable via effort, and stronger long-horizon agentic and research performance. Claude Fable 5 remains available at its original rate for existing integrations.
Yes, on both versions. Pass an image_url alongside your prompt and the model will reason about both the text and the image together — including charts, tables, PDFs, and screenshots.
Claude Fable 5.1 is $10.00 per million input tokens and $50.00 per million output tokens (Anthropic's own published rate), with a $0.0008 minimum per call. Claude Fable 5 is $8.00 per million input tokens and $40.00 per million output tokens, with a $0.001 minimum.
POST /api/v1/claude-fable-5-1 (or /claude-fable-5) is asynchronous: it returns a request_id you poll via /api/v1/predictions/{id}/result. Adding /stream to either path returns Server-Sent Events chunks in real time, ending in a data: [DONE] line, for live chat interfaces.
No. Extended thinking is always enabled on both Claude Fable versions and cannot be disabled, but its depth is adaptive — harder, multistep requests automatically use more reasoning effort.
No. Muapi exposes Claude Fable as standard REST endpoints. Create a Muapi API key, submit JSON, and either poll the returned request ID or read the SSE stream.
Create one Muapi API key and connect Anthropic's flagship reasoning models to your product through a plain REST call or a live SSE stream.