(Image credit: Future)

Claude Fable 5.1 arrived this week and it's Anthropic’s latest high-end model, built for complex tasks that require sustained reasoning. Its biggest focus is agentic work, meaning it can plan, use tools, check its own progress and keep working through problems with less supervision.With a 1-million-token context window, it's useful for huge codebases, lengthy documents and research-heavy projects. Anthropic says it delivers major gains in coding and long-running tasks while reducing the cost of some agentic workloads through cheaper caching. The simplest way to think about Fable 5.1 is that it’s designed to be given a project, not just a prompt.For those new to Claude who want to dive directly into Fable 5.1, here's the one prompt I suggest you try first.Give Fable 5.1 a project

(Image credit: Getty Images)Fable 5.1 can handle more, which means you'll want to give it a sturdier and more meaningful prompt.Here's a good prompt to see what Fable 5.1 can do: I'm building [describe a real project you're actually working on, in detail]. Here's my current approach: [paste your code, outline, or plan]. Tell me what's wrong with my thinking, what I'm missing, and then refactor or redesign the whole thing — but explain your reasoning at each decision point so I can learn from it.While this prompt may seem deceptively simple, it makes sense for an AI that can do so much more than previous models. I especially like it because you're not starting with a blank slate. Instead, you're giving the AI something of yours that you've already started. That's always a recipe for a better and more thorough response. Instead of asking Claude to create something from a clean set of instructions, you're handing it work that already exists, complete with your assumptions, mistakes and potentially questionable decisions. From there, Fable 5.1 has to figure out what you're trying to accomplish before deciding whether the way you're doing it actually makes sense.And crucially, you're explicitly giving Claude permission to tell you that you're wrong. One of the least useful things an AI assistant can do is take a bad premise and enthusiastically help you execute it. This prompt asks Fable 5.1 to interrogate the premise first. Only then does it get to rebuild the project.What the prompt is really testing