Claude Fable 5.1

Fable 5.1 is the model you reach for after Opus has tried and come up short. Anthropic placed it one tier above Opus in September 2026 for coding, research and problem solving that stretches over hours and many tool calls. It is the generally available form of Claude Mythos 5.1, with extra safeguards around biology and cybersecurity, and its biggest gains over Fable 5 are in agentic coding and long refactors.

What you can use it for

• Analysis and research where a mistake would be costly to undo
• Agents left alone for hours across a long chain of tools
• Front-end rewrites and refactors spread over a large repository
• Security review of code you own, for example hunting for vulnerabilities
• The leftover cases from your eval suite that cheaper models keep failing

Settings and limits

• Context window: 1,000,000 tokens; output up to 128K tokens per reply.
• Endpoints: /v1/chat/completions for the OpenAI SDK and /v1/messages for the Anthropic SDK and Claude Code.
• In Claude Code, Anthropic sets Fable 5.1 to High effort by default, so expect longer and more careful replies.
• Give the key a monthly spend cap in the dashboard before an unattended agent run.

Example request

from openai import OpenAI

client = OpenAI(
    base_url="https://kadraris.com/v1",
    api_key="sk-kd-...",
)
resp = client.chat.completions.create(
    model="claude-fable-5.1",
    messages=[{"role": "user",
               "content": "Trace why this payment job "
                          "double-charges under load."}],
)
print(resp.choices[0].message.content)

Swap the message for your own hardest bug. Setup for the CLI is in the Claude Code guide.

Questions

How do I keep Fable 5.1 spending under control?

The rate is $9.45 / $47.25 per 1M tokens, near the top of the chat price list. Check token counts on a few real tasks before routing a pipeline to it. A failed request costs $0.

When is Opus 5.5 the better call?

Most of the time. Anthropic itself says Claude Opus 5.5 reaches Fable 5.1 quality on the majority of work for less money. I would start there and move only the tasks that fail your tests. A side-by-side view is on the Fable 5.1 vs Opus 5.5 page.

Which Claude Code settings point it at Kadraris?

Set ANTHROPIC_BASE_URL to https://kadraris.com, use an sk-kd- key as the token and pick claude-fable-5.1 as the model name.

Claude Fable 5.1 in your code

from openai import OpenAI

client = OpenAI(base_url="https://kadraris.com/v1", api_key="sk-kd-...")
resp = client.chat.completions.create(
    model="claude-fable-5.1",
    messages=[{"role": "user", "content": "Suggest three opening shots for a short film about a cafe."}],
)
print(resp.choices[0].message.content)

curl https://kadraris.com/v1/chat/completions \
  -H "Authorization: Bearer $KADRARIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-fable-5.1", "messages": [{"role": "user", "content": "Suggest three opening shots for a short film about a cafe."}]}'

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://kadraris.com/v1", apiKey: process.env.KADRARIS_API_KEY });
const resp = await client.chat.completions.create({
  model: "claude-fable-5.1",
  messages: [{ role: "user", content: "Suggest three opening shots for a short film about a cafe." }],
});
console.log(resp.choices[0].message.content);

export ANTHROPIC_BASE_URL=https://kadraris.com
export ANTHROPIC_AUTH_TOKEN=sk-kd-...
export ANTHROPIC_MODEL=claude-fable-5.1
claude