Claude Models
Stop Using the Wrong Claude Model.
Most people are wasting their Claude usage before they even type a word. Here's how to match the model to the job.
Back to resourcesWhat Most People Get Wrong
Claude is not one model. It is a family of models, each built for a different kind of work.
The mistake is assuming the most powerful model is the right model for everything. It is not. Using Opus to write a summary is like using a chainsaw to slice bread. Technically it works. It is not the right tool.
Matching the model to the task is one of the fastest ways to get better output and use your Claude subscription more effectively.
The Models And What They Are Actually For
Haiku
Fast, cheap, high volume.
Haiku is not trying to be deep. It is built for repetitive tasks at scale: sorting and summarising large amounts of content, powering customer-facing chatbots, automating workflows that run constantly in the background. When you need speed and volume over nuance, Haiku is the model.
Sonnet
The one most people should be using.
Sonnet is fast and genuinely smart. It outperformed the older Opus on most benchmarks. For day-to-day work, emails, writing, summaries, coding, data tasks, brainstorming, Sonnet is your go-to. Most of the time, Sonnet is enough.
Opus
When the problem is genuinely hard.
Complex strategy, multi-step reasoning, creative work that needs real nuance. Opus is slower and costs more. That is worth it when the task actually demands it. The key word is genuinely. If Sonnet can handle it, use Sonnet.
Fable
Built for execution.
Building apps, creating agents, running complex workflows, tackling large projects with lots of moving parts. Fable is a specialist. The mistake people make is treating it as the best model. It is not the most capable across the board. It is the right tool for a specific kind of job.
Extended Thinking
Available on Opus, this is for problems you are stuck on.
When you turn it on, Claude spends more time reasoning before it responds. Not every problem needs it. When you are facing something genuinely complex with no clear path forward, it is the right tool.
How To Choose
Match the model to the task, not to what feels most impressive.
- Automation and high-volume tasks -> Haiku
- Day-to-day productivity, content, coding, data work -> Sonnet
- Complex strategy, deep analysis, hard decisions -> Opus
- Building apps, agents, or running workflows -> Fable
- Stuck on a problem with real nuance -> Opus with Extended Thinking
When in doubt, start with Sonnet. It will handle more than you expect.
The Cheat Sheet
Use This Before You Choose
- Haiku - Fast, cheap, high volume. Repetitive tasks, chatbots, automation.
- Sonnet - Fast and smart. The everyday model for most professionals.
- Opus - Slower, more powerful. For genuinely complex work.
- Fable - Built for execution. Apps, agents, and large multi-step projects.
- Extended Thinking - For problems that need more reasoning before a response.