Claude Fable 5.1, Explained Simply

Anthropic just released a new AI model designed for heavy-duty tasks like writing lots of code, doing deep research, and organizing huge amounts of information. It is really meant for developers, programmers, and people building apps who need an AI that can work on long, complicated projects without giving up halfway through.

What it does well

This model is built for "agentic" work. That means instead of just answering a single quick question, it can work on a multi-step task on its own, like researching a topic and writing a report, or writing and testing a big chunk of software. It also uses "adaptive thinking," which basically means the AI takes a moment to reason through a problem step-by-step before it answers, helping it make fewer mistakes on tricky tasks.

How it compares

It is a direct upgrade to Anthropic's previous model, Claude Fable 5. It is smarter and better at handling those long-running tasks, but it keeps the exact same price tag as the older version. Anthropic also released a sister model at the same time called Claude Mythos 5.1, but that one is specifically reserved for people in a special program called Project Glasswing.

What it costs and what it can handle

• Context window: 1 million tokens. A "token" is just a piece of a word, so this means it can read about 750,000 words at once—roughly the size of a few thick novels.
• Max output: 128,000 tokens. It can write incredibly long answers or huge blocks of code in a single go.
• Price: It costs $10 per million tokens of text you send in, and $50 per million tokens it writes out.
• Cache reads: $0.25 per million tokens. "Caching" is when the AI remembers text you just showed it so it doesn't have to re-read everything from scratch, making it much cheaper and faster.

Worth knowing

You cannot just chat with this model on a normal consumer website. It is only available through developer tools (APIs) and cloud platforms like Amazon Web Services, Google Cloud, and Microsoft Foundry. Also, while it is the same price as the older model to use normally, you only get the big money savings on "cache reads" if you are set up to use that specific memory feature.