claude-fable-5-2

Claude Fable 5.2: reasoning for demanding technical work.

API IDclaude-fable-5-2
View setup
View usage

Estimate your cost

Explore the cost of your next build.

Estimate text chat and coding. Other model types are not included yet.

Start with an example or enter your own

Tokens measure text. Input includes your prompt, instructions and history. Output includes generated text and billable reasoning. Use an example if unsure.

YOUR ESTIMATEUSD
Claude Fable 5.2
Estimated monthly cost
$48.00/ month

100 requests/day × 30 days = 3,000 requests/month

Per request
$0.016
Input / month
$12.00
Output / month
$36.00

Pay as you go · Published discounts included

Prompt cachingOff

Shares of total input, averaged across requests. The rest uses regular input pricing. Include cold starts, cache expiry and writes in your assumptions. Published rates do not guarantee cache hits; eligibility and setup vary by model. Uses the published write rate only; one-hour cache writes are excluded.

Compare models

Same token volumes and cache shares for each model. Missing required cache rates make an estimate unavailable. Actual output lengths and quality can differ.

Have a budget?

Approximately 625 requests at this request size.About 6.3 days at 100 requests/day.

Text only · Caching off · Media and tool fees excluded.
Agent tasks can make multiple requests. Final costs appear in Usage.

Overview

Context
1,000,000 tokens
Max output
16,384 tokens
Input
Text
Output
Text
Providers
Anthropic Foundry Endpoint
Latest operational check
Healthy
API compatibility
  • OpenAI Chat Completions
  • Anthropic Messages
Recommended for
  • General-purpose reasoning
  • Long-context analysis
  • Production assistants

Keep in mind: Adaptive thinking is always on; omit sampling controls. Output limit is 16K. Images, hosted tools and long-context requests are unavailable.

Supported features
  • Function calling: Supported
  • Structured outputs: Not supported
  • Vision: Not supported
  • Streaming: Supported
  • Reasoning: Supported
  • Prompt caching: Supported
Supported parameters
toolsstream
View API reference

Make your first request

View full setup
curl --fail-with-body https://gateway.primexos.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-fable-5-2","messages":[{"role":"user","content":"Hello from Primex"}],"stream":false}'

Executing this request uses your account’s API credits.