Claude Fable 5.2 Is Here: What's New and Can You Use It?
Anthropic's Fable 5.2 is faster and 20% cheaper than Fable 5.1. It isn't on the free plan, but Pro and Team users can try it at no extra cost until 16 November.

In this post5 sections
Anthropic released Claude Fable 5.2 on 9 November 2026. It's the company's most capable model that the public can use, and Anthropic says it is up to 30% faster than Fable 5.1 while costing 20% less for developers. If you pay for Claude, you can try it this week, and Pro and Team users get a free trial window that ends on 16 November.
The catch is that it isn't on the free plan. There's also a more powerful sister model, Mythos 5.2, that most people will never get to touch. Here's who gets what, and whether it's worth caring about.
What Claude Fable 5.2 does better
Fable is built for hard, drawn-out technical work, not quick chats. According to Anthropic's announcement, it is strongest at coding across big projects, maths, science, finance and design work such as 3D and hardware. Amazon's AWS launch post adds that it checks its own work, catches its own mistakes and can keep going on a task for hours or even days.
It's also a reasoning model, which means it thinks through a problem before it answers. Anthropic says the default "medium effort" setting in the Claude apps uses fewer tokens (the small chunks of text that AI usage is measured in) than Fable 5.1 did. In plain terms: you get the same thinking without burning through your limits as fast.
Anthropic also says the safety filters for cyber and biology topics now block about 55% fewer requests than in Fable 5.1. If Fable 5.2 still refuses something in those areas, your question gets handed to an older Claude model instead.
Who should actually use it? Anthropic's product lead Dianne Penn told CNBC that Fable is meant for "power users or power use cases", and that Claude Opus 5.5 is the better fit for everyday tasks. If you mostly use Claude for emails, summaries and planning, Opus 5.5 is still the sensible default.
Is Claude Fable 5.2 free? No, but there's a one-week offer
Anthropic's help page on Fable plans puts it plainly: "Fable models aren't available on the Free plan." Some coverage says Fable 5.2 is open to "all users", but that only means it isn't restricted the way Mythos is. You still need a paid plan.
Here's how access works on each plan, based on the official launch offer page:
- Pro and standard Team seats: from 9 November until 11:59 pm PT on 16 November (1:29 pm IST on 17 November), Fable usage comes out of your normal weekly limit, and up to 50% of that limit can go to Fable. You also get one free limit reset that expires at the same time.
- After the offer ends: Pro and standard Team users can keep using Fable only through paid usage credits, which are charged at API rates, and only if you have credits switched on.
- Max, Ultra and premium seats: Fable is already included, up to 50% of weekly limits. You get one to three resets depending on how long you've been paying, valid until 8 March 2027.
- Not included in the offer: the Free plan, plus some Enterprise plan types. On Enterprise, an admin has to switch Fable 5.2 on.
One warning from Anthropic: Fable eats through your limits faster than other Claude models, so 50% of your week goes quicker than it sounds.
Mythos 5.2 is the same brain, kept behind a locked door
Anthropic says Fable 5.2 and Claude Mythos 5.2 are the same underlying model. Fable just has stronger safeguards around cybersecurity and biology. Mythos 5.2 is only for organisations vetted through Anthropic's Cyber Verification and Life Sciences Verification programmes, because it is very good at finding software security holes.
In Anthropic's tests, Mythos 5.2 found 10.2% more vulnerabilities than Opus 5.5. Mozilla's Brian Grinstead said it "roughly doubled the vulnerability discovery metrics" of Mythos 5.1. For regular users, the takeaway is simple: you can't sign up for Mythos, and you don't need it.
Claude vs ChatGPT: Fable leads on most of Anthropic's own tests
Anthropic compared Fable 5.2 with OpenAI's GPT-6 Astra and Google's Gemini 4 Carbon. Keep in mind these are Anthropic's own numbers, not independent tests:
- FrontierSWE v2 (coding): Fable 5.2 69.9%, GPT-6 Astra 65.5%, Gemini 4 Carbon 61.2%
- Humanity's Last Exam, with tools (very hard questions): Fable 5.2 74.2%, GPT-6 Astra 57.2%
- OSWorld 2.1 (using a computer by itself): Fable 5.2 89.3%, GPT-6 Astra 79.5%
- AutomationBench: Gemini 4 Carbon leads at 53.9%, ahead of Fable 5.2 at 51.4%
On price for developers, Fable 5.2 costs $8 per million input tokens and $40 per million output tokens. OpenAI's pricing page lists GPT-6 Astra at $10 and $50. Fable is slightly cheaper on paper, but the real cost depends on how many tokens each model uses to finish a job. Anthropic hasn't given Indian rupee pricing for Fable usage credits.
For chat app users the question is simpler. Neither company puts its top model on the free plan, so if you want the best from either one, you'll be paying.
What to do before 16 November
If you already pay for Claude Pro, this week is the cheapest way to see whether Fable 5.2 is worth it to you. Give it your hardest real problem, like a messy spreadsheet, a coding bug or a long research question, and use your free reset before it expires. Coders using Claude Code will need version 2.1.321 or later.
If you're on the free plan, don't upgrade just for Fable unless you do heavy technical work. On Pro it stops being included after the trial week. Want better answers from whichever model you use? Our AI Prompt Playbook covers the prompts that help most. And keep an eye on Anthropic: CNBC reports the company could go public as soon as next week.
Sources
Found this useful? Send it to someone who should read it.


