What is Claude Fable 5.2?
Claude Fable 5.2 is Anthropic's model for people who need to send in a lot and get a lot back. It succeeds Fable 5.1 and pairs built-in reasoning with a very large context window and long outputs. It is sold in the Claude app on paid plans and through the API, with a premium fast mode for latency-sensitive work. The official model page gives its release date as November 9, 2026.
What is new in Claude Fable 5.2?
The headline change is speed. Anthropic says the model generates outputs up to 30% faster than Claude Fable 5.1. That is a ceiling, so your results will vary by task. The second change is how the model reasons: Anthropic says Fable 5.2 has always-on adaptive thinking. In general terms, adaptive thinking means the model decides how much reasoning a request needs instead of applying a fixed amount. A simple question should get a quick answer, and a hard one more deliberation. "Always-on" means this is not a mode you enable, so you do not have to decide per request.
Capacity is the third piece. The model has a 1M-token context window and a maximum output of 128K tokens. That lets you send a large codebase or a stack of documents in one request and get a long result back without splitting the job into parts.
Who can use Claude Fable 5.2, and where?
In the Claude app, Fable 5.2 is available to Pro, Max, Ultra, Team, and Enterprise users. Developers get broader access: the model is available to all customers on the Claude API.
You can also reach it through cloud platforms. Anthropic's release notes list it on Claude Platform on AWS and several other cloud platforms. That helps if your team already buys infrastructure through a cloud provider and wants usage on the same invoice.
How much does Claude Fable 5.2 cost?
Standard API pricing is $8 per million input tokens and $40 per million output tokens. Output costs five times as much as input, so long generations, such as responses approaching the 128K-token maximum, will drive your bill faster than large prompts will.
There is also a premium tier. Fast mode costs $16 per million input tokens and $80 per million output tokens, double the standard rates on both sides. Consider it for interactive features where waiting is the main cost. For batch jobs and background work, standard pricing is the sensible default.
If you run Fable 5.1 today, measure latency on your own prompts before switching. The speed gain is an upper bound, and your workload decides how much of it you see.
What can you use today?
Claude Fable 5.2 is available on Imagera today.
- LLM Arena: send one prompt to several current language models and compare the answers side by side.
- Imagera Code: a coding agent that works in your project with the model you pick.
Frequently Asked Questions
Where can I find the full technical specifications?
Start with the Claude Fable 5.2 overview in Anthropic's developer documentation. The what's new page covers what changed for API customers.
Where can I read Anthropic's own announcement?
Anthropic published it as Introducing Claude Fable 5.2 and Claude Mythos 5.2. Platform changes are tracked in the Claude Platform release notes.



