Claude Code lets you pick which AI model runs your code tasks, and you change it in one place

When you use Claude Code — the AI coding assistant built into Claude — it defaults to Claude 3.5 Sonnet, which is fast and handles most coding work well. If you need a different model because a task is too complex, too straightforward, or you want to test something specific, you can switch models without restarting or losing your work. The change happens in your settings, and it takes effect on your next request.

Which model you pick affects how much processing power your computer uses while Claude thinks through the problem. Smaller models run faster and use less memory. Larger models are slower but can handle more complicated code tasks. You do not need to understand the technical details — you just need to know that switching models is a choice you make based on what you are trying to do.

Key Takeaways

  • You change your Claude model in the settings menu at the bottom left of the Claude interface, where it shows the current model name.
  • Claude 3.5 Sonnet is the default and works for most coding tasks without slowing down your computer noticeably.
  • Claude 3 Opus is larger and slower but handles very complex code problems better if Sonnet struggles.
  • Switching models does not affect code you have already written — it only changes which model Claude uses for new requests.
  • Your choice of model stays in place until you change it again, so you can test different models for the same task.

Where to find the model selector

Open Claude in your browser and look at the bottom left corner of the screen. You will see a button or dropdown that displays the name of the model you are currently using — usually "Claude 3.5 Sonnet" when you first start. Click on that button or dropdown arrow next to it.

A menu appears showing all the models available to you. The list depends on your Claude subscription level. Free users see fewer options than paid subscribers. Select the model you want to use, and Claude switches to it when ready. You do not need to refresh the page or close any windows.

Understanding the three main models and when to use each

Claude 3.5 Sonnet is the middle ground. It is fast enough that you will not notice lag on most computers, and it handles the majority of coding tasks — debugging, writing functions, refactoring, explaining code. It is the default because it balances speed and capability well. Use this for everyday coding work.

Claude 3 Opus is the largest and slowest model. It thinks longer about hard problems and can handle very complex code tasks, like designing a system architecture or untangling deeply nested logic. The trade-off is that it takes longer to respond, and your computer has to work harder while Claude processes the request. Use this when Sonnet gives you an answer that does not quite work, or when you are tackling something you know is complicated.

Claude 3 Haiku is the smallest and fastest. It is good for straightforward tasks like fixing a typo in code, explaining what a short function does, or generating boilerplate. It uses less of your computer's resources. Use this when you know the task is straightforward and you want the answer in seconds rather than waiting.

How switching models affects your computer's performance

When you switch to a larger model like Opus, your computer's CPU and memory usage goes up while Claude is thinking. You might notice your fan running harder, other programs slowing down slightly, or the response taking 30 seconds to a minute instead of 10 seconds. This is normal — it does not mean something is broken.

If your computer is already running slowly or you have many programs open, switching to Opus might make things feel sluggish. In that case, close other programs first, or stick with Sonnet. Haiku uses the least resources, so if you are on an older computer or a laptop running on battery, Haiku is the gentler choice.

The model you choose does not permanently change how your computer runs. As soon as Claude finishes responding, your computer goes back to normal. Switching back to Sonnet or Haiku when ready frees up those resources.

Testing different models for the same problem

You can ask Claude the same question using different models to see which one gives you the best answer. Type your question, get a response from Sonnet, then switch to Opus and ask the same thing. Compare the two answers. This is useful when Sonnet's answer does not quite solve your problem, or when you want to see if a smaller model can handle the task.

Your conversation history stays the same when you switch models — Claude can see everything you have already discussed. So you can say "that did not work, try a different approach" after switching, and Claude understands the context.

What happens to your code when you switch models

Switching models does not change or delete any code you have already written or generated. The code stays exactly as it is. The model switch only affects what Claude generates next. If you have written 50 lines of code with Sonnet and then switch to Opus, those 50 lines do not change — only new code Claude writes will come from Opus.

This means you can safely experiment. Switch to a different model, ask Claude to rewrite a section, and if you do not like the result, you can switch back and ask Sonnet to try again. You are not locked into one model for a whole conversation.

When you might not have all models available

Free Claude users have access to Claude 3.5 Sonnet and sometimes Claude 3 Haiku, but not Opus. Paid subscribers (Claude Pro) have access to all models. If you click the model selector and see only one or two options, that is why — your account level limits which models you can use.

If you need Opus for a specific task and you do not have it, you have two choices: upgrade to Claude Pro, or break the task into smaller pieces that Sonnet can handle. Many complex problems can be solved by asking Sonnet to work through them step by step, even if it takes a few back-and-forth exchanges.

Frequently Asked Questions

Does switching models cost me more money?

Claude Pro is a flat monthly subscription, so switching between models does not change your bill. Free users do not pay anything, but they have fewer models to choose from. If you are a free user and want access to Opus, you would need to subscribe to Claude Pro.

Will switching to a bigger model make my computer crash?

No. Larger models use more resources, but Claude is designed to handle this safely. Your computer might slow down temporarily while Claude is thinking, but it will not crash. If you are worried, close other programs first to free up memory.

Can I set a default model so I do not have to switch every time?

Yes. Once you select a model, Claude remembers it and uses it for all future conversations until you change it again. You do not have to pick a model each time you start a new chat.

What if I switch models in the middle of a conversation?

Claude keeps your entire conversation history and can see everything you have discussed so far. The new model picks up where the old one left off and understands the context. You can switch back and forth as many times as you want in a single conversation.

How do I know which model is best for my task?

Start with Sonnet — it works for most things. If the answer is not good enough or Claude says the task is too hard, try Opus. If the task is very straightforward and you want a fast answer, try Haiku. After a few tries, you will get a feel for which model works best for your typical coding problems.