All resources Tools and Setup
A free guide by ZAK
Which AI Model to Use for Each Kind of Work
Model names change all the time. The choice underneath them does not. Here is how I decide which AI model to use, fast, standard or deep thinking, so you get good answers without waiting or overpaying.
Open any AI tool and you will see a model picker, and often an effort or thinking setting too. Most people leave it on whatever was there.
That works for a lot of tasks. But matching the model to the job saves time on small tasks and gets you much better answers on hard ones.
Step 1
Know the Three Tiers
Every major tool now offers some version of these, whatever the names are this month.
- Fast models. Quick answers, low cost. Great for rewording, short summaries and simple questions.
- Standard models. The everyday default. Good writing, solid reasoning, fast enough.
- Thinking or reasoning models. Take longer and reason in stages before they reply. Best for analysis, planning, numbers and tricky decisions.
Step 2
Match the Tier to the Task
Use this as a quick guide.
- Rewrite an email or summarise a page | fast
- Draft a report, a post or a plan | standard
- Compare options, find flaws, work with numbers | thinking
- Long documents or many files | a model with a large context window
- Anything legal, financial or safety related | thinking, then a qualified person
Step 3
Turn Up Effort Only When It Pays
Many tools let you set effort or thinking time. Higher effort reads more, checks more and takes longer. Use it for decisions with real money or risk attached, and leave it lower for drafting and quick questions.
Step 4
Test Your Own Top Three Tasks
Run your three most common tasks on a fast and a deep model. If the fast one is good enough, use it by default.
I am going to run the same task on two models. Score each answer 1 to 5 on accuracy, depth and usefulness, and tell me which you would pick for this task in future and why. Task: [PASTE] Answer A: [PASTE] Answer B: [PASTE]
Questions About Which AI model to use
Which AI model should I use?
Use a fast model for quick tasks, a standard model for everyday writing and planning and a thinking model for analysis and big decisions.
What is a reasoning or thinking model?
A model that spends extra time reasoning in stages before it replies, which helps on complex tasks.
Does a bigger model always give better answers?
No. For simple tasks a fast model is often just as good and quicker. Test on your own work.
What do AI effort settings change?
Higher effort lets the model read and reason more before answering, which takes longer and may cost more.
Next step
Put This Guide to Work
Bring the jobs your team runs most and we can match each one to the model that suits it.