TL;DR
What is Claude Fable 5.2?
Anthropic released Claude Fable 5.2, its most capable public AI model, on Monday. It costs 20% less to build with than Fable 5.1 and answers up to 30% faster. If your paid Claude plan includes Fable, you get extra Fable usage this week, so try it on your hardest job now.
What happened
Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2 on Monday, November 9, 2026, and posted the launch on X at 1:57pm ET. Anthropic says both models improve on coding, reasoning and vision, explain themselves more clearly, and write their answers up to 30% faster than the 5.1 versions.
Fable and Mythos are the same underlying model. Fable ships with strong cybersecurity and biology safeguards, so anyone can use it. Mythos drops some of those blocks and goes only to organizations approved through Anthropic's Cyber and Life Sciences Verification Programs.
What changed from Fable 5.1
- Lower price. US$8 per million input tokens and US$40 per million output tokens, down from US$10 and US$50. Tokens are the chunks of text AI is billed by.
- Faster answers. Up to 30% faster output than Fable 5.1, per Anthropic.
- Optional fast mode. Up to 2.5 times faster output, at double the price.
- Fewer false blocks. Anthropic counted about 55% fewer unnecessary safety blocks than Fable 5.1 in early use across claude.ai, Claude Code and Cowork. When a request is blocked, Claude answers with an older Opus model instead.
- Same big memory. A 1M-token context window (how much text the model can read at once) and up to 128,000 tokens of output.
What it costs
Prices are per million tokens, in US dollars, before tax, from Anthropic's model page. Fable 5.2 and Mythos 5.2 share the same prices.
| Option | Input | Output |
|---|---|---|
| Fable 5.1 (old price) | US$10 | US$50 |
| Fable 5.2, standard | US$8 | US$40 |
| Fable 5.2, batch (results later, not instantly) | US$4 | US$20 |
| Fable 5.2, fast mode | US$16 | US$80 |
TechCrunch reports the cut makes Fable 5.2 cheaper than its main rival, OpenAI's GPT-6 Astra.
Who can use it
| Who | What you get |
|---|---|
| Paid Claude plans that include Fable | Extra Fable usage this week: one to three Fable resets, based on how long you've subscribed. On Pro and standard Team seats, Fable can use up to 50% of weekly limits. |
| Developers on the Claude API | Fable 5.2 now, as claude-fable-5-2 |
| Amazon, Google Cloud and Microsoft customers | Fable 5.2 on Amazon Bedrock, Google Cloud and Microsoft Foundry |
| Verified security and life sciences groups | Mythos 5.2, after approval through Anthropic's verification programs |
| Claude Enterprise customers | Claude Security, Anthropic's code scanner, now runs on models that include Mythos 5.2 |
Anthropic's own advice: for most coding and knowledge work, start with Claude Opus 5.5, its "strongest daily driver," and move up to Fable only for the biggest, most open-ended jobs.
For ways to spend less on Claude plans, read our Claude Pro alternatives.
What Anthropic says it can do
Anthropic says Fable 5.2 solved two open problems in black-hole physics and helped design three working electronic devices. It also says a team of about 1,000 Fable 5.2 agents formalized the proof of Fermat's Last Theorem, turning it into a computer-checkable proof, in 30 hours. All three are Anthropic's own claims about its own work.
On benchmarks, Anthropic's table puts Fable 5.2 ahead of Opus 5.5 on almost all measures, though Google's Gemini 4 Carbon scored higher on the AutomationBench business-workflow benchmark (53.9% vs 51.4%).
For developers: check before you switch
Anthropic lists four changes that can break existing code:
- Editing earlier turns of a conversation and sending Fable 5.2's saved reasoning back now returns an error on every account.
- Only Fable 5.2 and Mythos 5.2 can read Fable 5.2's saved reasoning.
- The older
computer_20251124computer-use tool returns an error everywhere except Amazon Bedrock. - With the advisor tool, a Fable 5.2 executor needs a Fable 5.2 or Mythos 5.2 advisor.
If Claude Code, claude.ai or the Claude Agent SDK builds your requests, the first change doesn't affect you. Fable 5.2 also requires 30-day data retention unless Anthropic approves otherwise.
X reactions
Shortcut, which now offers Fable 5.2 in its app, posted its own numbers within the hour:
Fable 5.2 is now available in Shortcut!
— Shortcut (@tryshortcutai) November 9, 2026
Fable 5.2 sets a new accuracy record on our financial modeling benchmarks, scoring 91% at high effort versus 86% for Fable 5.1, and its low effort beats every other model's best score.
Our take
A 20% price cut on a top model is real money for anyone running it all day. Just don't read "cheaper per token" as "cheaper per job." In his post on X, Zapier CEO Wade Foster said Fable 5.2 averaged about US$4.64 per task at max effort on Zapier's workflows, against US$1.28 for Opus 5.5, because it makes more tool calls and checks its work. Thorough costs extra.
So use this week's bonus usage the smart way. Give Fable the one job that keeps beating your usual model, like a messy spreadsheet or a long project plan, and see if the result earns the bigger bill. For everyday writing, email and quick questions, Anthropic itself says Opus 5.5 is the place to start.
— Artificially Nerdy
What to do this week
- Paid Claude subscribers: pick Fable 5.2 for one hard task while the extra usage lasts.
- API builders: run your four-point check above before changing the model ID to
claude-fable-5-2. - Cost-watchers: compare cost per finished job, not just per token, before moving everyday work to Fable.
- Security and research teams: apply through Anthropic's Cyber or Life Sciences Verification Program if you need Mythos 5.2.
FAQ
Is Claude Fable 5.2 free?
No. It's on paid Claude plans that include Fable, and on the API at US$8 per million input tokens and US$40 per million output tokens.
What's the difference between Fable 5.2 and Mythos 5.2?
They're the same model. Fable keeps strong cybersecurity and biology safeguards and is open to everyone; Mythos goes only to organizations Anthropic has verified for security or life sciences work.
How long does the extra Fable usage last?
Anthropic says "this week." It gives eligible plans one to three Fable resets based on how long you've subscribed; it hasn't published a per-user amount beyond that.
Should I switch everything from Opus 5.5 to Fable 5.2?
Anthropic doesn't recommend it. It suggests starting with Opus 5.5 for most tasks and moving to Fable when a job is too big or open-ended.
Sources
- Anthropic: Claude Fable 5.2 and Mythos 5.2 — November 9, 2026. Launch, speed, prices, fast mode, safeguards, research claims, extra usage, availability.
- Claude Platform docs: What's new in Claude Fable 5.2 — price table, old price, batch price, context window, breaking changes, platforms, data retention.
- Claude Platform release notes — November 9, 2026 entry.
- TechCrunch: Anthropic releases new Mythos and Fable models — November 9, 2026. Price comparison with GPT-6 Astra.
- Claude on X: launch post — November 9, 2026, 1:57pm ET.
- Wade Foster on X — November 9, 2026, 2:27pm ET. Zapier results and per-task costs.
- Shortcut on X — November 9, 2026, 2:16pm ET.
