Wanna understand how to make your AI do its best thinking for you, without using up all your tokens? It’s pretty easy once you understand what to do.
EFFORT
Depending on the platform, you may see settings such as low, medium, high, extra high, or something similar. The labels vary, but the basic idea is the same. You’re telling the AI how much effort it should put into working through the problem.
That does not mean higher is always better. I mean, it is, but it also burns through your tokens.
For routine work, lower thinking settings are usually plenty. Rewriting a paragraph, summarizing something straightforward, or cleaning up an email can generally be done with low or medium effort very well.
MODEL
Chose a higher model (e.g., Sol vs. Terra OR Opus vs. Sonnet) when the work gets more complicated and requires better, more creative, thinking. For strategy, analysis, comparing options, reviewing a complicated document, finding flaws in an argument, or working through several competing constraints all benefit from giving the model more room to reason.
The simplest rule is to match the effort to the job. Start lower for ordinary work. Raise it when the task is long, complicated, ambiguous, or requires real judgment.
And if you turn the thinking all the way up for every request, the AI police won’t arrest you. You’ll just spend more time and tokens asking a very sophisticated system to think deeply about whether “Thanks!” needs an exclamation point.