Model ID: claude-fable-5
Claude Fable5 API access
From zero to a working claude-fable-5 call in four steps — plus ready-to-run code in Python, TypeScript, and cURL, and every official channel compared. No reseller required, ever.
Get access in four steps
Fable5 is generally available — any funded Claude Console account can call it.
01
Create a Claude Console account
Sign up at the Claude Console and create an API key. New accounts get a one-time starter credit (about $5) that works with any model, including Fable5 — no credit card needed to try it.
02
Call model ID claude-fable-5
Point your existing Messages API call at claude-fable-5 — the request shape is unchanged from other Claude models. The ID is already a pinned snapshot, so there is no dated variant to add and no version to chase.
03
Set the effort parameter
temperature and top_p are gone. Use effort (low → xhigh) to trade reasoning depth against latency and cost; high is the right default for engineering work.
04
Control costs from day one
Enable prompt caching (90% off cache hits) and route asynchronous work through the Batch API (50% off). Cache-heavy agents land at 1.3–1.6x Opus cost instead of the 2x sticker price.
Code examples
The same Messages API as every Claude model — only the model string and the effort parameter are new.
import anthropic
client = anthropic.Anthropic() # reads ANTHROPIC_API_KEY
message = client.messages.create(
model="claude-fable-5",
max_tokens=2048,
effort="high", # low | medium | high | xhigh
system="You are a precise coding assistant.",
messages=[{"role": "user", "content": "Refactor this function..."}],
)
print(message.content[0].text)Fable5 follows instructions literally — your system prompt matters more than before. Write it with the prompt guide
Every official channel
Four ways to reach the same model, as of June 10, 2026.
Claude API (direct)
Lowest latency and first to get new features. $10/$50 per million tokens, with caching and Batch discounts available.
OpenRouter
Listed as anthropic/claude-fable-5. One integration surface for A/B testing Fable5 against GPT-5.5, Gemini 3.1 Pro, and others.
AWS · Google Cloud · Microsoft Foundry
All three cloud platforms serve Fable5 — useful when procurement, compliance, or data residency rules out a direct Anthropic account.
GitHub Copilot
Generally available inside Copilot — no separate Anthropic key needed; billing goes through your Copilot plan.
You never need a third-party reseller. Be cautious of anyone selling Fable5 access keys outside these official channels.
Check the full spec sheetAPI FAQ
Quick answers on Claude Fable5 API access.
Build on Claude Fable5 with your eyes open
Know the Claude Fable5 cost math, the safeguard behavior, and the right tier before you commit. Create a free account to get notified when specs, prices, or access channels change.