Anthropic released two new models on 9 November 2026: Claude Fable 5.2, which anyone on a paid plan or the API can use, and Claude Mythos 5.2, which only verified organizations can use. This explainer covers what changed, how Fable and Mythos differ, what it costs, and how to try Fable 5.2 today in the Claude apps, Claude Code and the API. Details are accurate as of 14 November 2026.
What happened
Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2 as upgrades to its top-tier models, aimed at the most ambitious, open-ended technical work. They are upgrades to the 5.1 versions.
According to the announcement, the two models “show improvements in areas like coding, reasoning, and vision,” and “generate outputs up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1.”
Anthropic’s developer docs describe Fable 5.2 as built “for demanding reasoning and long-horizon agentic work,” meaning things like agent sessions that run for hours, multistep research, and analysis that ends in a finished document, spreadsheet or deck.
Fable vs Mythos: same model, different access
Fable and Mythos are not two separate brains. Anthropic says both are the same underlying model. The difference is safeguards and who can use it.
- Fable 5.2 includes cybersecurity and biology safeguards, so it is open to all API customers and to paid Claude plans. Classifiers block high-risk requests such as exploit generation and most research biology tasks. When a query is blocked, Anthropic says users “will instead receive a response from Opus 4.8 for cybersecurity and Opus 5 for biology.”
- Mythos 5.2 is the same model without some of those safeguards. It is limited to organizations verified through Anthropic’s programs. “Cybersecurity professionals can apply for access to Claude Mythos 5.2 through the Cyber Verification Program,” and life sciences researchers can apply through the Life Sciences Verification Program.
Anthropic also says Fable 5.2 blocks legitimate work less often: “about 55% fewer blocks” than Fable 5.1 overall, and “about 60% fewer in cybersecurity and 40% fewer in biology.”
What changed from Fable 5.1
Here is what we know from official pages:
- Better and faster. Gains in coding, reasoning and vision, with output up to 30% faster.
- Cheaper. Fable 5.1 cost $10 per million input tokens and $50 per million output tokens. Fable 5.2 costs $8 and $40.
- Benchmarks. Anthropic says that on its FrontierCode 2.0 coding test, “At Xhigh effort, Fable 5.2 reaches 49.1%, the highest score of any released model we’ve tested.” For security work, “Claude Mythos 5.2 found 10.2% more vulnerabilities than Claude Opus 5.5” on its CVE Rediscovery test. These are Anthropic’s own results, not independent tests.
- Same size limits. A 1M-token context window and up to 128K output tokens.
For developers there are also some breaking changes, covered in What’s new in Claude Fable 5.2. The main one: if your code edits earlier turns of a conversation, the API can now return a 400 error.
What we don’t know yet: how Fable 5.2 performs in independent, third-party testing, and how the launch promotions on consumer plans will settle once launch week ends.
Pricing
From Anthropic’s official pricing page, as of 14 November 2026:
- Standard: $8 per million input tokens, $40 per million output tokens (same for Mythos 5.2).
- Prompt caching: cache reads cost $0.20 per million tokens. Cache writes are $10 (5-minute) or $16 (1-hour).
- Batch API: 50% off, so $4 input and $20 output.
- Fast mode: $16 input and $80 output per million tokens, for up to 2.5x higher output speed. Fast mode doesn’t work with the Batch API.
For context, Claude Opus 5.5 is $4 and $20, half the price of Fable 5.2.
How to try Fable 5.2 today
In the Claude apps
Anthropic’s help article on Fable models says Fable 5.2 is available on all paid plans (Pro, Max, Ultra, Team and Enterprise). On the web, desktop or mobile app, choose Fable 5.2 from the model picker. Menu layouts can change.
How it’s billed depends on your plan:
- Max, Ultra and premium Team or seat-based Enterprise seats: included, using up to 50% of your weekly usage limits.
- Pro and standard Team or Enterprise seats: runs on usage credits.
- Usage-based Enterprise: billed at API rates.
The announcement also mentions extra Fable usage for eligible paid subscribers during launch week. See the announcement for the promotion details. On Enterprise plans, an admin has to switch Fable 5.2 on.
In Claude Code
Fable 5.2 needs Claude Code v2.1.321 or later. Update, then pick the model:
claude update
claude --model fable
Inside a session you can also run /model fable. Claude Code defaults Fable 5.2 to medium effort. Change it with /effort high or similar. To turn on fast mode, run /fast. On subscription plans, fast mode always bills to usage credits. If you are new to Claude Code, our guide to Agent Skills in Claude Code is a good next step.
With the API
The model ID is claude-fable-5-2. You need an API key from the Claude Console, set as ANTHROPIC_API_KEY. This request follows the shape of Anthropic’s current docs:
curl https://api.anthropic.com/v1/messages \
-H "x-api-key: $ANTHROPIC_API_KEY" \
-H "anthropic-version: 2026-10-21" \
-H "content-type: application/json" \
-d '{
"model": "claude-fable-5-2",
"max_tokens": 32000,
"messages": [{
"role": "user",
"content": "Explain the trade-offs between microservices and a monolith"
}],
"output_config": {
"effort": "high"
}
}'
And the same request in Python (pip install anthropic first):
import anthropic
client = anthropic.Anthropic()
response = client.messages.create(
model="claude-fable-5-2",
max_tokens=32000,
messages=[
{"role": "user", "content": "Explain the trade-offs between microservices and a monolith"}
],
output_config={"effort": "high"},
)
for block in response.content:
if block.type == "text":
print(block.text)
On the API the default effort is high. Thinking is always on and counts toward max_tokens, so Anthropic recommends a large value at high and above. To try fast mode, add "speed": "fast" to the request.
On cloud platforms, the ID is global.anthropic.claude-fable-5-2 on Amazon Bedrock and claude-fable-5-2 on Google Cloud, Microsoft Foundry and Claude Platform on AWS. Fast mode works on the Claude API and Google Cloud, and on Anthropic-hosted Foundry deployments through a separate claude-fable-5-2-fast deployment (without the speed field). It is not available on Bedrock.
Who should switch
Anthropic’s own advice is to start with Claude Opus 5.5 for most work. Move up to Fable 5.2 when Opus at higher effort still falls short. In practice:
- Switch if you run long agent tasks, hard coding or multistep research, or you already use Fable 5.1. You get a cheaper, faster model.
- Stay put if you mostly write, summarize or chat. Opus 5.5 or Sonnet 5.5 costs less.
- Teams rolling out AI tools should apply the same planning they would for any workplace assistant. Our ChatGPT Business setup guide covers the admin questions worth asking.
Common problems and fixes
- 400 error about thinking. Fable 5.2 doesn’t accept
thinking: {"type": "disabled"}or abudget_tokenssetting. Leavethinkingout and use effort instead. - 400 error about sampling. Leave out
temperature,top_pandtop_k. - 400 error with
tool_choice. Forced tool use (anyor a named tool) isn’t supported. Useauto. - A reply with
stop_reason: "refusal". A safety classifier blocked the request. You can set up fallback to another model, as described in the refusals docs. - Fable missing in Claude Code. Update to v2.1.321 or later. If the picker says “Requires usage credits,” your plan bills Fable to credits.
- Not visible at work. On Team and Enterprise, an admin controls which models are available.
Common questions
Is Fable 5.2 free? No. It requires a paid Claude plan or API billing.
Can I get Mythos 5.2? Only through one of Anthropic’s verification programs, such as the Cyber Verification Program or the Life Sciences Verification Program.
Does my Fable 5.1 code still work? Usually after changing the model ID, but review the breaking changes in the migration guide first.
Wrap-up
Fable 5.2 is a faster, cheaper and, by Anthropic’s measures, stronger version of its top model, with Mythos 5.2 as the same model for vetted security and science teams. Most people don’t need it for everyday work. If you do hard, long-running technical tasks, it’s worth a test.
Teams using Claude together can also try Claude Whiteboards, a shared board for laying out ideas with Claude. Our step-by-step guide covers who can use it and how to turn it on.
Sources
Links last checked on 15 November 2026.
- Anthropic: Introducing Claude Fable 5.2 and Claude Mythos 5.2
- Claude Docs: Models overview
- Claude Docs: Claude Fable 5.2
- Claude Docs: What’s new in Claude Fable 5.2
- Claude Docs: Pricing
- Claude Docs: Fast mode
- Claude Docs: Effort
- Claude Code Docs: Model configuration
- Claude Code Docs: Fast mode
- Claude Help Center: Claude Fable models on your plan
- Claude Help Center: Cyber Verification Program





Write a response