Claude Fable 5.2 API

Fable 5.2 tops Anthropic’s line-up. It is for work where a mistake is expensive: redesigning a module, untangling a legacy codebase of hundreds of files, writing a migration you cannot roll back. In Claude Code it keeps the plan through long sessions and sees large tasks through to the end.

It costs less than Fable 5.1: input and output are 20% cheaper ($0.8 and $4 per 1M here vs $1 and $5), and so are cache reads ($0.02 vs $0.025). Claude Code and Cursor resend the same project context all the time, so the saving shows over a long session.

Anthropic’s models for code, writing and agents. Haiku is fast and cheap, Sonnet is the everyday workhorse, Opus and Fable take the hardest tasks. The natural pick for Claude Code.

The model answers in the Anthropic Messages format and in the OpenAI Chat Completions format. Only the address and the key change.

curl https://api.klinch.cloud/v1/messages \
  -H "x-api-key: kl_live_…" \
  -H "content-type: application/json" \
  -d '{"model": "claude-fable-5-2", "max_tokens": 1024,
       "messages": [{"role": "user", "content": "Hello"}]}'

from openai import OpenAI

client = OpenAI(base_url="https://api.klinch.cloud/v1", api_key="kl_live_…")
r = client.chat.completions.create(
    model="claude-fable-5-2",
    messages=[{"role": "user", "content": "Hello"}],
)
print(r.choices[0].message.content)

API docs