# Claude: Which Model to Use for Which Job in Your Business Published 12 November 2026 by Prad Patel. Tags: AI Tools, Small Business. Page: https://kaizenaiconsulting.com/claude-which-model-for-which-job/ Claude now comes in four models, and the model menu in the Claude app is easy to ignore. Most people pick one on day one and never touch it again. That costs them either speed or usage allowance, and sometimes both. On Wednesday, Anthropic's own Claude account posted a short guide to choosing between them. We have turned it into the version we would give a client: which model to use for which everyday job in a small business, and what you trade each time. The short answer: start on **Opus 5.5**, drop to **Sonnet 5.5** when you want faster replies or are running low on usage, step up to **Fable 5.2** for the big, open-ended questions, and leave **Haiku 5.5** to the automations that run behind the scenes. ## What Claude itself recommends Here is the guidance, in Anthropic's words, from [Claude's post on LinkedIn on 11 November](https://www.linkedin.com/posts/claude_heres-how-to-choose-between-the-available-activity-7526290254320840705-xNtX): > "In general, use Opus 5.5 as your daily driver for coding and knowledge work. Switch to Sonnet 5.5 if you want faster replies and more runway with your usage. For your most ambitious, open-ended technical and research work, reach for Fable 5.2. If you're building on the API, Haiku 5.5 is our fastest and lightest model for high-volume work behind the scenes. These are rough guidelines, not rules: the best model is the one that fits the way you work." That last line matters. There is no wrong answer here, only a better fit. Anthropic's [developer guide to choosing a model](https://platform.claude.com/docs/en/about-claude/models/choosing-a-model) says the same thing in more technical language: start most work on Opus and move to Fable only when Opus is genuinely falling short. Fable 5.2 is also brand new. It was [released on 9 November](https://9to5mac.com/2026/11/09/anthropic-upgrades-claude-with-new-fable-5-2-model-details-here/), replacing Fable 5.1 as Anthropic's most capable model, so if you last looked at the model menu a few weeks ago, it has changed. ## Opus 5.5: your everyday model **Use it for:** writing a quote, answering a tricky client email, checking a contract or terms and conditions for anything unusual, turning meeting notes into a proposal, sorting out a spreadsheet. This is the model we would leave selected by default. It is clever enough for nearly everything a small business throws at it, and it handles long documents well. If you only ever use one model, make it this one. **The trade-off:** it is not the fastest, and it uses more of your plan's allowance than Sonnet. On a busy day of back-and-forth, that is when you will notice. ## Sonnet 5.5: faster replies, more runway **Use it for:** quick replies, rewording a message so it sounds friendlier, social media captions, product descriptions, a first draft you are going to edit anyway, and anything where you are going back and forth many times. Claude's own advice is to switch to Sonnet "if Opus is breezing through your work". In plain English: if the job is not hard enough to need the bigger model, you are spending allowance you do not need to spend. Sonnet is strong on most tasks and replies faster, which matters more than you would think when you are working through a pile of small jobs. **The trade-off:** on a genuinely complicated job (a long contract, a messy set of figures, a tricky judgement call) you may need a second attempt. If you find yourself correcting it, switch back to Opus. ## Fable 5.2: the big, open-ended questions **Use it for:** the work you would normally block out an afternoon for. Researching a new market or supplier, building a business plan, comparing five software options against your actual needs, or working through a long project that takes many steps. Fable is Anthropic's most capable model, built for what it calls "long-horizon" work: tasks that need sustained thinking over many steps rather than a single answer. For most owners that will be a handful of times a month, not every day. **The trade-off:** it is the slowest of the four and the heaviest on your usage. Using it for a quick email is like hiring a consultant to post a letter. Save it for the questions that deserve it. ## Haiku 5.5: the one working behind the scenes **Use it for:** automations rather than conversations. Sorting hundreds of website enquiries into "quote", "complaint" and "general", tagging incoming invoices, pulling the key details out of booking forms, or drafting first replies that a person then checks. Claude's post is clear that Haiku is aimed at people "building on the API", meaning tools and automations that call Claude directly, rather than the chat app. It is the fastest and lightest model, which is exactly what you want when the same small job runs hundreds of times a day. This is the model we reach for when we build automations for clients with tools like n8n. **The trade-off:** it is not the one to ask for advice or judgement. Give it simple, repeatable jobs with clear instructions. ## A simple rule you can stick on the wall - **Default:** Opus 5.5. - **Opus breezing through it, or you are low on usage?** Sonnet 5.5. - **Big, ambitious, open-ended question?** Fable 5.2. - **Running the same job hundreds of times in an automation?** Haiku 5.5. Switching takes two clicks: in the Claude app, open the model menu beside the message box and pick the one you want. You can change it mid-conversation, so you can do the heavy thinking on Opus or Fable and then switch to Sonnet for the quick follow-ups. ## What this could do for you The benefit is not a magic number, it is fewer of the two frustrations people tell us about most: hitting the usage limit halfway through the afternoon, and waiting for a long answer to a short question. Matching the model to the job helps with both. Sonnet for the quick stuff means your allowance lasts longer for the work that needs Opus. Fable for the occasional big question means you get a deeper answer the first time rather than three rounds of "try again". It is worth getting right because so many owners are now in the app every day. [Simply Business's 2026 SME Insights Report](https://www.simplybusiness.co.uk/knowledge/business-news/small-business-ai-usage-doubles/) found 47% of UK small business owners now use AI tools, more than double the 22% a year earlier. If that includes you, five minutes spent learning the model menu pays back every week. ## One thing to check: your plan Which models you can pick depends on your Claude plan, and Anthropic has been changing its plans this month. At the time of writing, the free plan does not include Opus, and Fable uses up a paid plan's allowance faster than the others. Check [Claude's pricing page](https://claude.com/pricing) for what your plan includes today before you rely on a particular model. If you would like help setting Claude up for the way your business actually works, from the right plan to the right model for each job, book a free 30 minute call with us. **Prad Patel** ran Glass Tailors for twelve years, growing it to 12 staff and around £1m turnover. He now builds AI systems, quote tools and marketing for owner-run businesses, and shows them how they work. [About Prad](https://kaizenaiconsulting.com/about/)