Claude Fable 5.1: AI Large Language Models
Generate text with Claude Fable 5.1. Anthropic's flagship model with adaptive thinking, 1M context. Try free — pay per generation, no subscription.
Generate text with Claude Fable 5.1. Anthropic's flagship model with adaptive thinking, 1M context. Try free — pay per generation, no subscription.
About this model
Claude Fable 5.1 is Anthropic's newest flagship model, built for long-horizon agentic coding, multistep research, and document, spreadsheet, and slide work. It runs with adaptive extended thinking always enabled (steerable via effort, from low up to max) and a 1 million token context window with up to 128K tokens of output. It accepts both text and image inputs — charts, tables, PDFs, and screenshots — for multimodal analysis, and is a step up in reasoning depth from the earlier Claude Fable 5. Pricing is token-based: $10.00 per million input tokens and $50.00 per million output tokens.
Cost analysis
| Provider | Cost | Notes |
|---|---|---|
| muapiapp | $10.00/M 输入 Token, $50.00/M 输出 Token | Token-based billing at Anthropic's official published rate. Minimum $0.0008 per call. |
| Fal.ai | Niedostępne | Claude Fable 5.1 is 不可用 on Fal.ai. |
| Replicate | Niedostępne | Claude Fable 5.1 is 不可用 on Replicate. |
Token-based billing at Anthropic's official published rate. Minimum $0.0008 per call.
Claude Fable 5.1 is 不可用 on Fal.ai.
Claude Fable 5.1 is 不可用 on Replicate.
** Competitor pricing is estimated based on similar model architectures and usage tiers.
Configuration schema
| Parameter | Type | Description | Default |
|---|---|---|---|
| Prompt | string | The user message or instruction. | Explain quantum entanglement in simple terms. |
| Adres URL obrazu | string | Optional image URL for multimodal requests. | undefined |
| System Prompt | string | Optional system-level instruction to guide model behavior. | You are a concise and precise assistant. |
The user message or instruction.
Explain quantum entanglement in simple terms.Optional image URL for multimodal requests.
undefinedOptional system-level instruction to guide model behavior.
You are a concise and precise assistant.Developer documentation
POST /api/v1/claude-fable-5-1 — returns request_id, poll for the result via /api/v1/predictions/{id}/result.
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": "Summarize the key risks in this quarterly report and list follow-up questions.",
"image_url": "https://example.com/report-page-1.png",
"system_prompt": "You are a meticulous financial analyst."
}'
POST /api/v1/claude-fable-5-1/stream — returns a live Server-Sent Events stream (Content-Type: text/event-stream). Each chunk has the format data: {"choices":[{"delta":{"content":"text"}}]}, ending with data: [DONE].
See Streaming Documentation for full code examples.
Frequently asked
Claude Fable 5.1 is Anthropic's newest flagship language model, built for long-horizon agentic coding, multistep research, and document, spreadsheet, and slide analysis, with a 1M-token context window and always-on adaptive extended thinking.
Pricing is token-based: $10.00 per million input tokens and $50.00 per million output tokens. The minimum charge per call is $0.0008. Actual cost is deducted after each call based on token counts returned by the model.
/claude-fable-5-1 is async — you receive a request_id and poll for the result. /claude-fable-5-1/stream returns a live SSE stream. Use streaming for chat UIs; use the async endpoint for workflows and automation.
Yes. Pass an image_url in the request body to include an image — such as a chart, table, PDF page, or screenshot — as part of the user message for multimodal analysis.
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 compared to [Claude Fable 5](/playground/claude-fable-5).
Extended thinking is always enabled and cannot be turned off, but its depth is adaptive to the task — harder, multistep requests automatically get more reasoning effort.