Claude Fable 5.2
The newest version of Anthropic’s top tier, for demanding reasoning and long, multi-step work, available anonymized through NeverNo without an Anthropic account.
Updated November 10, 2026
Readable on your devices only.
Your conversation is readable on your device, and on your other devices if you sync them. Nowhere else.
Unreadable to us.
Your message passes through without being written down; your chats are only ever stored encrypted.
The model doesn’t know you.
It gets the question, never your email, your account or your IP address.
Nothing is left along the way.
The reply comes back and is stored here, encrypted; even when you sync, your key never reaches us in the clear.
At a glance
| Maker | Anthropic |
|---|---|
| Context | 200,000 tokens |
| Longest reply | 16,384 tokens |
| Strengths in the app | Hard questions, Reasoning, Code, Reads images |
| Reasoning | Always on |
| Who can use it | Paid plans |
| How it’s paid | Credits, from the first use |
| Privacy | Anonymized: NeverNo sends the request to its maker without your name, account or IP address. The maker still reads the text, applies its own rules and may keep it for a while under its own policy.Some requests its maker doesn’t allow are answered by one of NeverNo’s private models instead, and the app tells you so. |
Price
What the app charges, read from the same table the servers bill with. 1 credit = $0.01.
| Typical message (2,000 tokens read, 600 written) | 10.55 credits ($0.11) |
|---|---|
| Million tokens read | 2,110 credits ($21.10) |
| Million tokens written | 10,550 credits ($105.50) |
Measured
Our own measurements and one third-party score, each with its conditions and date. Figures are not rounded.
| Median time to the first word | 7.97 sOur client, 12 everyday requests run 3 times each, November 11, 2026 |
|---|
What it is
Claude Fable 5.2 is the newest version of the model at the top of Anthropic’s range. It has been offered since November 9, 2026 and replaced Claude Fable 5.1 on NeverNo the next day; if you had picked 5.1, your choice now goes to 5.2 on its own. When we read Anthropic’s models overview on October 31, 2026, the company described Fable as its model “for demanding reasoning and long-horizon agentic work”. On NeverNo it is the model to bring in when a task has many moving parts, when a mistake halfway through would be expensive, or when you need an answer you can defend line by line.
NeverNo gives you access to it in anonymized form. Your request reaches Anthropic through NeverNo, without your name, your account or your IP address, and your conversation history stays encrypted in your browser. Anthropic still reads the text it receives, applies its own usage rules and may keep requests for a while under its own policy; Fable is therefore not a private model in the sense of our open models, and it may decline some requests. Some requests its maker does not allow are answered by one of NeverNo’s private models instead, and the app tells you so.
Fable 5.2 always reasons before answering, and that reasoning cannot be turned down. In our probe of November 10, 2026, the settings that ask for no reasoning were rejected, the ones that ask for a smaller effort had no effect, and with our real instructions it still thought for several hundred tokens before writing. There is therefore no level to choose in the selector. It accepts 200,000 tokens of context, writes up to 16,384 tokens per answer, reads images and is tagged in the selector for reasoning and for code.
The most visible change with this version is speed. In our own test of 11 November 2026 (12 everyday requests, three runs each, timed from our client), its median time to the first word was 7.97 seconds, versus 1.27 seconds for Fable 5.1 on the same bench, which puts it on the lowest bar of the speed gauge in the selector. It remains a premium model on paid plans, paid in credits, and the most expensive of the four Claude models on NeverNo: worth it for the hard cases, wasteful for small talk.
Strengths
Built by Anthropic for its most demanding work: long chains of reasoning, plans with many dependencies and problems that need checking at each step.
Made for code as much as for reasoning: tagged for programming in the selector, with a 200k context to hold several files, logs and a specification together.
Reads images, so architecture diagrams, screenshots of errors or photographed pages can be part of the problem.
Anonymized: no Anthropic account, no identity sent, and the conversation stored encrypted on your own device.
Limits
It is the costliest Claude model on NeverNo and uses credits on every message; Claude Sonnet 5.5 or Claude Opus 5.5 cover most needs for less.
It is slow to start: 7.97 seconds median before the first word in our 11 November 2026 test, because its reasoning cannot be shortened, even for a one-line question. Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 all answered sooner on our bench.
As an anonymized model, it is subject to the rules of Anthropic, which reads the text and may refuse; some requests outside those rules are handed to a private NeverNo model, which the app signals.
Answers are capped at 16,384 tokens each: a very large refactor or report comes in several passes.
When to choose it
Debugging and refactoring across a whole project
Paste the relevant files, the failing test output and what the code should do. Fable can trace a bug across modules, propose a fix, then explain which other parts of the project the change touches.
Planning a complex piece of work
A migration, a product launch, a research protocol: give it the constraints, the deadlines and the unknowns, and ask for an ordered plan with risks and checkpoints. It is designed for exactly this kind of long-horizon thinking.
Second opinion on a high-stakes analysis
Submit a financial model, a legal argument or a technical design and ask it to find the weakest assumption. Its step-by-step reasoning makes its objections easy to verify before you act on them.
Prompt tips
Give it the whole situation in one go: goal, constraints, what you have tried, what success looks like. Fable does its best work when it can reason over the full picture, and since every turn starts with a few seconds of thinking, one complete message beats five short ones.
Ask it to state its assumptions before concluding, or to list what it would check next; you will see where its reasoning could go wrong and can correct it early.
Keep it for the hard part of a task. Draft with a lighter model, then hand Fable the tricky step or the final review, which keeps your credit spend proportional.
Frequently asked questions
Can I use Claude Fable 5.2 for free?
No. On NeverNo, Claude Fable 5.2 is a premium model on paid plans, paid in credits per use. An estimate for a typical exchange is shown in the model selector before you send.
Is Claude Fable 5.2 private on NeverNo?
It is anonymized: Anthropic receives the text without your name, account or IP address, and your history stays encrypted in your browser. Anthropic still reads the request, applies its own rules and may keep it for a while under its policy. For conversations that never reach a closed company, use one of NeverNo’s private models.
What changed between Claude Fable 5.1 and 5.2?
Fable 5.2 has been offered since November 9, 2026 and replaced 5.1 on NeverNo the next day; a choice of 5.1 now goes to 5.2. The change you will notice is speed: 7.97 seconds median before the first word in our test of 11 November 2026, versus 1.27 seconds for 5.1, because 5.2 always reasons and that cannot be turned down. On NeverNo, the context (200,000 tokens), the answer length (up to 16,384 tokens) and image reading stay the same.
Claude Fable 5.2 or Claude Opus 5.5: what is the difference?
Anthropic positions Fable as its top tier for demanding reasoning and long, agentic tasks, and it is tagged for code in our selector; Opus 5.5 costs fewer credits and started answering much sooner in our tests (2.35 seconds median against 7.97). For long, difficult projects, Fable; for demanding writing, analysis and conversation, Opus is often enough.
Do I need an Anthropic account to use Claude Fable 5.2?
No. With a paid NeverNo plan you choose Claude Fable 5.2 in the model selector and write as usual: NeverNo forwards the request without your identity, and there is no separate Anthropic subscription to manage.