Stage 2 · Lesson 5 of 6
When a bigger model pays for itself
A bigger model pays off when the cost of a wrong answer is higher than the extra price. Work it out. Say a costs 0.1 cents per request and is right 90% of the time, and a large one costs 1 cent and is right 98% of the time.
If each wrong answer costs you a human fix worth 20 cents, the small model loses about 2 cents per request in fixes and the large model about 0.4 cents. The large model wins. If a wrong answer costs nothing (a suggestion the user ignores), the small one wins easily.
This simple sum, price plus expected cost of errors, settles most "is it worth it?" arguments better than any .
A question to check yourself, an AI practice session and 1 cheat sheet entry come with the full roadmap. Unlock
Resources for this stage
Start with· follow one of these
- CoursePaidGrokking Prompt Engineering for Software Engineers
Design Gurus · Beginner · Subscription
Write prompts with clear rules and formats, so you can test instruction following properly
- VideoFreeDeep Dive into Large Language Models like ChatGPT
Andrej Karpathy · Intermediate
Free: a clear explanation of how reasoning models are trained and why thinking helps
Practise
Keep handy
Also worth your time
How reasoning effort works in practice and where the reasoning tokens show up in usage
- Large Language Model CourseGo deeper
Free: go deeper into how models are trained, which explains where reasoning comes from