Model guides

Claude Sonnet 5.5 online: what it is for and how to use it

Anthropic built Sonnet 5.5 for fast, well-scoped coding, writing and document work. Here is where to use it in a browser, a worked example and when to switch models.

On this page
  1. What Anthropic built Sonnet 5.5 for
  2. Where you can use Sonnet 5.5 online
  3. How Sonnet 5.5 appears in Chat AI App
  4. Worked example: fix a bug and rewrite the email it feeds
  5. Check the output before you use it
  6. When to switch to another model
  7. Use Claude Sonnet 5.5 in Chat AI App
  8. Common questions
  9. Sources

You can use Claude Sonnet 5.5 online in Anthropic’s own Claude apps, or as one model in a multi-model app such as Chat AI App. Anthropic describes Sonnet 5.5 as a faster, lower-cost complement to Claude Opus 5.5. Use it for well-scoped writing, coding and document tasks, and switch models when the work becomes open-ended.

This guide covers what Anthropic says the model is for and how it appears in Chat AI App’s model picker. It then walks through a writing and code example and the signs that another model fits the job better.

What Anthropic built Sonnet 5.5 for

Anthropic released Claude Sonnet 5.5 on September 28, 2026. Its model overview positions it as the balance of speed and intelligence in the current Claude lineup. It accepts text and images and returns text.

The launch announcement says Sonnet 5.5 is strongest at well-scoped everyday tasks, fixing bugs, and creating polished documents, slides and spreadsheets. It also says the model writes more clearly than Anthropic’s previous generation. Anthropic’s Sonnet product page lists four areas of use:

  • Coding: everyday development, such as building features, fixing bugs and reviewing code.
  • Agents: well-defined tasks such as investigation, review and drafting.
  • Design and visual work: polishing interfaces, creating diagrams and user flows, and building slides from templates.
  • Enterprise workflows: knowledge work and office tasks such as analysis, documents and slides, with clearer and more direct writing.

Anthropic is also clear about the boundary. The same announcement says Opus 5.5 remains clearly stronger at complex, open-ended work that requires sustained judgment. Treat all of this as the vendor’s description, not as a reason to skip checking the output.

Where you can use Sonnet 5.5 online

There are three common routes. They suit different people, so compare them by what you need around the model.

RouteWho it suitsWhat to check
Claude.ai and the Claude appsPeople who want Anthropic’s own app. Anthropic says anyone can chat with Sonnet 5.5 on Claude.ai, on web, iOS and Android.Anthropic’s plans and usage terms for the features you need.
Claude API and cloud platformsDevelopers building their own product. Anthropic lists the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry.You write the integration and pay for usage under that platform’s terms.
Chat AI AppPeople who want Sonnet 5.5 next to GPT, Gemini, Grok and other models in one conversation, with file uploads and downloads.It is a separate, independent subscription. Features of Anthropic’s apps do not carry over.

The deciding question is whether you want one model family or several. Chat AI App is built for the second case: several model families, your files and downloadable documents under one subscription.

How Sonnet 5.5 appears in Chat AI App

In Chat AI App, the Claude lineup is available through the picker's Claude choices, Claude Opus 5.5 and Claude Fable 5.1. Pick it yourself when you want that model for a task. If you would rather not choose, AI Chat (Auto) chooses a model for your request.

The picker also lists Claude Opus 5.5 and Claude Fable 5.1, plus GPT-6, Gemini 3.1 Pro, Grok 4.20, DeepSeek V4 and others. You can switch models mid-conversation without starting over. That makes it practical to draft with Sonnet 5.5 and ask a different model to review the same thread.

A model label tells you which picker option you chose. It does not bring Anthropic’s app features, account settings or usage rules with it. Chat AI App is an independent product and is not affiliated with Anthropic.

Worked example: fix a bug and rewrite the email it feeds

Fictional example: Marcus maintains the booking site for Spoke & Chain, a made-up bike repair shop. Customers in Chicago get confirmation emails that show the wrong pickup day. A booking for Tuesday, November 3, 2026 shows as Monday, November 2. The email text is also wordy and buries the cancellation rule.

This is a well-scoped job with a clear finish, the kind of task Anthropic describes Sonnet 5.5 as suited to. Marcus pastes the function and the email draft, then sends this prompt:

Help me with two linked tasks. Task 1: the JavaScript function below formats the pickup date in our confirmation email. For the input “2026-11-03”, customers in Chicago see “Mon, Nov 2”. The expected output is “Tue, Nov 3”. Explain the cause in two sentences. Give the smallest fix and keep the function name and arguments. Then list three test inputs with expected outputs. Task 2: rewrite the email below in under 120 words. Keep the shop name, pickup address and cancellation rule exactly as written. Do not add discounts, apologies or promises. Code: [paste function]. Email: [paste draft].

A useful answer names the cause. A date-only string such as “2026-11-03” is read as midnight UTC, which is still the previous evening in Chicago. The fix should build the date from its parts or format it in UTC. Run the three tests yourself, including one for a timezone east of UTC.

For the email, compare the rewrite with the original line by line. The cancellation rule must survive word for word. If the draft drifts, send one narrow correction: “Restore the cancellation rule exactly as written. Change nothing else.”

Check the output before you use it

  • Run the tests in two timezones: Run the three tests with TZ=America/Chicago and again with TZ=Asia/Tokyo. Both runs must print “Tue, Nov 3” for “2026-11-03”.
  • Read the diff: Confirm the change is as small as you asked. Anthropic’s Sonnet 5.5 prompting guide notes that in coding work it tends to add tests or small supporting files you did not ask for.
  • Remove what you did not ask for: If the answer adds a helper file or a new test framework, ask Sonnet 5.5 to fold the fix back into the one function. Ask it to list each line it changed.
  • Diff the email: The shop name, pickup address and cancellation rule must match the original character for character.
  • Check current details: The same guide says the model sometimes answers from training knowledge when a search would catch changes. Ask for a web search when rules or prices may have moved.

When to switch to another model

Sonnet 5.5 is a sensible default for scoped work. Switch when the task changes shape:

  • The work turns open-ended: Anthropic’s prompting guide says an Opus model is the better choice for the hardest long-horizon work. Try Claude Opus 5.5 for a long analysis across several documents.
  • You need the most demanding reasoning: Anthropic’s model selection guide points to Claude Fable 5.1 for its highest available capability. It appears in the picker too.
  • You need current facts with sources: Pick Sonar or Sonar Pro (Perplexity web search), or ask for a web search and follow the sources.
  • You want a second opinion: Switch to GPT-6 or Gemini 3.1 Pro in the same conversation. Ask it to review the draft against your checklist.
  • You need an image: Use Image Generation rather than a text model.

Keep any comparison fair. Give each model the same brief and judge the results against the same checklist. One impressive answer is a weak basis for choosing a daily tool.

Use Claude Sonnet 5.5 in Chat AI App

  1. Open AI Chat and choose a Claude model in the model picker. Pick AI Chat (Auto) instead if you would rather not choose.
  2. Paste your code and draft, or attach files with + → Files. You can attach PDFs, Word documents, spreadsheets, slides and images, and several files in one message.
  3. Send a prompt with a clear finish, like the example above. Keep follow-ups in the same conversation so the constraints stay visible.
  4. Switch models from the picker when the job grows. You do not need to start a new conversation.
  5. Ask for a Word, Excel, PowerPoint or PDF file when you need one, and use the reply’s Download controls. Check the formatting after download.

Plans cost $29.95 per month, $66.50 per quarter or $179.00 per year, and you can cancel anytime. See current pricing and terms before you subscribe. To see every model in one place, browse our model guides.

Common questions

Can I use Claude Sonnet 5.5 online without an Anthropic subscription?
Anthropic says anyone can chat with Sonnet 5.5 on Claude.ai. In Chat AI App, Claude models are included under your Chat AI App plan. No separate Anthropic subscription is needed to pick it.
Is Chat AI App the same as Claude.ai?
No. Chat AI App is an independent product and is not affiliated with Anthropic. Claude models appear in its model picker alongside other model families. Features, settings and usage terms of Anthropic’s apps do not carry over.
Should I use Claude Sonnet 5.5 or Claude Opus 5.5?
Start with Sonnet 5.5 for well-scoped writing, coding and document tasks. Anthropic says Opus 5.5 remains stronger at complex, open-ended work, so switch when the task needs sustained judgment across a lot of material.
Can I switch away from Sonnet 5.5 in the middle of a chat?
Yes. You can switch models mid-conversation without starting over, for example to ask Claude Opus 5.5 or GPT-6 to review a draft that Sonnet 5.5 wrote.

Sources and further reading