Claude Fable 5.2 — agentic coding and sustained problem solving
Claude Fable 5.2 helps you review a migration before it reaches production. Work with text, images and PDFs in a 1M-token context. Compare current token rates below.
- Adaptive thinking
- 1M-token context
- Images and PDFs
- Prompt caching
- Streaming
Try Claude Fable 5.2
Chat with Claude Fable 5.2, attach images or PDFs, and see the tokens and cost of every answer.
claude-fable-5-2$2.80 in · $14.00 out per 1M tokensUnsupported sampling settings and unknown fields are ignored. Optional beta features require their documented header.
Try one of these prompts
Messages are billed from your balance at the rates below, like an API call. Failed requests are free.
A prompt to try with Claude Fable 5.2
Use this as a starting point, then add the evidence and constraints from your own project.
Help me review a migration before it reaches production. Separate observations from assumptions, identify the most consequential unknown, and propose a small test that could change the decision.
Suggested input, not a recorded model response.
Claude Fable 5.2 pricing
Claude Fable 5.2 is billed per token. Default uses the bare model ID; additional tiers appear alongside it when available, with their own callable IDs and rates.
Claude Fable 5.2 costs follow reported usage. Example token counts are estimates.
- Input
- $2.80Price · 1M tokens
- Output
- $14.00Price · 1M tokens
- Cache read
- $0.07Price · 1M tokens
- Cache write, 5 minutes
- $3.50Price · 1M tokens
- Cache write, 1 hour
- $5.60Price · 1M tokens
What is Claude Fable 5.2?
Claude Fable 5.2 is an Anthropic language model for agentic coding and sustained problem solving. It accepts text, images and PDFs in a 1M-token context and produces up to 128K tokens, including thinking. Bring source files or documents into the request to ground its answer.
Claude Fable 5.2 defaults to high effort with adaptive thinking. Use low for routine work or raise effort for harder questions. State your acceptance criteria and verify the result against the supplied evidence.
Model ID: claude-fable-5-2 · Input: text, images, PDFs · Output: text · Formats: Anthropic Messages, OpenAI Chat Completions, OpenAI Responses.
Two ways to use Claude Fable 5.2
Use Claude Fable 5.2 from your application or coding agent.
Call it from your backend
Point the Anthropic SDK at SeedRouter. Claude Fable 5.2 also accepts OpenAI formats.
- 1Create an API key
- 2Set the base URL to https://api.seedrouter.ai
- 3Send claude-fable-5-2 as the model
Hand it to your coding agent
Copy the prompt, then review the request and estimated cost in your agent.
- 1Copy the prompt
- 2Paste it into your agent
- 3Approve the request
What Claude Fable 5.2 is good at
Give Claude Fable 5.2 a concrete goal, source material and a way to check the result.
Long coding sessions
Have Claude Fable 5.2 trace a change across files and propose tests for the failure you are fixing.
Research and analysis
Ask Claude Fable 5.2 to compare supplied reports, cite source passages and flag missing evidence.
Documents and files
Claude Fable 5.2 reads PDFs, charts and screenshots alongside your text.
Agent workflows
Let Claude Fable 5.2 request tools while your application controls permissions and execution.
Effort you control
Start Claude Fable 5.2 at low effort for routine tasks; reserve max for difficult problems.
Cheaper repeat context
Reuse a stable prompt prefix with Claude Fable 5.2. Check separate cache read and write rates below.
Ways to put Claude Fable 5.2 to work
Code review bots
Give Claude Fable 5.2 a patch and surrounding code. Ask for reproducible failures and a validation plan.
Support assistants
Have Claude Fable 5.2 draft answers from product documentation and flag questions that need a person.
Document extraction
Ask Claude Fable 5.2 to extract document fields with a source passage beside each value.
Start using Claude Fable 5.2 in four steps
- 01
Create a key
Sign in and create an API key. Add credit whenever you need it; there is no subscription.
- 02
Change the base URL
Point the Anthropic or OpenAI SDK at https://api.seedrouter.ai. Your code stays the same.
- 03
Pick Claude Fable 5.2
Send claude-fable-5-2 as the model. Set effort to trade depth for speed and cost.
- 04
Check the bill
Every request shows its tokens and charge in your usage records.
Claude Fable 5.2 on SeedRouter at a glance
What you can use today.
- Anthropic Messages API
- Yes, official format
- OpenAI Chat Completions
- Yes
- OpenAI Responses
- Yes
- Image and PDF input
- Yes, by URL or base64
- Prompt caching
- 5-minute and 1-hour
- Forced tool use
- No; choose auto or none
- Failed requests
- Not charged
The native Messages format exposes model-specific controls. OpenAI-compatible formats use their own request schemas.
Claude Fable 5.2 limits
1M-token context
On Claude Fable 5.2, prompt, files and history share one 1M-token window.
128K tokens out
One Claude Fable 5.2 request can return up to 128K tokens, thinking included.
Thinking stays on
Claude Fable 5.2 keeps adaptive thinking on. Lower effort for routine tasks.
Sampling settings
Claude Fable 5.2 uses effort to control reasoning. Unsupported sampling settings are ignored; they do not make answers deterministic.
Why call Claude Fable 5.2 through SeedRouter
One key, three formats
Call Claude Fable 5.2 with Anthropic Messages, OpenAI Chat Completions or Responses, all with the same key.
Pay per token
Top up once and spend it on any model. No plan, no monthly fee.
Failures are free
Failed Claude Fable 5.2 requests are not charged.
Live prices
Compare Default and any additional tiers side by side. The selected tier determines the rate charged for that request.
Prompt caching
Put reusable material first in a Claude Fable 5.2 prompt. Changing the prefix can prevent cache hits.
Images and PDFs
Give Claude Fable 5.2 charts, screenshots or PDFs to interpret.
Streaming
Stream Claude Fable 5.2 tokens as they arrive with stream: true.
Related models
Other models you can call with the same key.
Call Claude Fable 5 with the Anthropic or OpenAI SDK. Deep reasoning, 1M-token context, forced tool use, live per-token pricing and a playground to test it.
View pricingCall Claude Fable 5.1 with the Anthropic or OpenAI SDK. Deep reasoning, 1M-token context, cheap cached reads, live pricing and a playground to try it today.
View pricingTry Claude Haiku 5.5 for classification, extraction and summaries. See live token prices, test your prompts, and connect your app with the Messages API.
View pricingTry Claude Mythos 5 for complex reasoning and research with a 1M-token context. Compare live token prices and model tiers, then test it in the playground.
View pricingTry Claude Mythos 5.1 for research and agent workflows with a 1M-token context. Compare live token prices and model tiers, then test it in the playground.
View pricingCall Claude Opus 5.5 with the Anthropic or OpenAI SDK. 1M-token context, adaptive thinking, live per-token pricing and a playground you can try right now.
View pricingClaude Fable 5.2 guides
Walkthroughs and comparisons for this model.
Guides are on the way.
Claude Fable 5.2 FAQ
What is Claude Fable 5.2?+
Claude Fable 5.2 is an Anthropic model for agentic coding and sustained problem solving. It supports a 1M-token context and up to 128K output tokens, including thinking.
When should I choose Claude Fable 5.2?+
Test Claude Fable 5.2 on work you can evaluate, then compare correctness and total charge with Claude Opus 5.5.
How much does Claude Fable 5.2 cost?+
Claude Fable 5.2 costs $2.80 / 1M input and $14.00 / 1M output per million tokens on Default. The live table lists cache rates and any additional tiers.
What is the default effort of Claude Fable 5.2?+
Claude Fable 5.2 defaults to high. Set output_config.effort to low, medium, high, xhigh or max. Thinking stays adaptive; fixed thinking budgets are unsupported.
When was Claude Fable 5.2 released and what is its knowledge cutoff?+
Claude Fable 5.2 launched November 9, 2026, with a June 2026 knowledge cutoff. Supply current sources for recent events.
What is the Claude Fable 5.2 model ID and how do I call it?+
Use claude-fable-5-2 with your SeedRouter key. Set the Anthropic SDK base URL to https://api.seedrouter.ai and send model, max_tokens and messages.
Does Claude Fable 5.2 work with the OpenAI SDK?+
Yes. Claude Fable 5.2 accepts Chat Completions and Responses at https://api.seedrouter.ai/v1 with the same key and model ID.
Am I charged if a Claude Fable 5.2 request fails?+
Failed Claude Fable 5.2 requests are not charged. A successful answer stopped at its output limit is charged for reported usage.
How do I choose a Claude Fable 5.2 pricing tier?+
Use claude-fable-5-2 for Default. For another available tier, copy its full claude-fable-5-2:group ID. Prices and permitted choices appear in the playground.
Try Claude Fable 5.2 now
Send your first Claude Fable 5.2 request from the playground or your own code in minutes.