Lesson 67 · 2 min
Small models and large models
Bigger models know more and reason better, while smaller ones are faster, cheaper and can run on your own device.
A model's size is roughly its number of parameters, the learned numbers inside it. Large models have hundreds of billions or more. Small ones may have a few billion, small enough for a laptop or phone.
Large models are better at hard reasoning, rare facts, long documents and tricky writing. They run in data centers, so they cost more per answer and take a bit longer.
Small models are quick and cheap, and some run without the internet, which keeps your data on your device. They handle everyday jobs well: sorting messages, short summaries, simple rewrites, autocomplete.
Many apps quietly use both: a small model for easy requests and a large one when the question is hard. For your own use, try the small or fast option first and step up only when the answer is not good enough.
Do it now
Find the smallest model that works
- 1. Many chatbots let you pick a fast mode and a more capable mode.
- 2. Give both the same everyday task, like the prompt below.
- 3. If the fast answer is good enough, use it for that kind of task from now on.
Rewrite this message so it is polite and clear, under 60 words: "hey, the report is late again, need it by friday or we have a problem"
Quick check
Keep going
Next →Context window vs memoryWords from this lesson
- Small language model
- A compact language model that is cheaper and faster, often good enough for focused tasks and able to run on devices.
- Parameters
- The adjustable numbers inside a model that are set during training. Big models have billions of them.
- LLM
- An AI trained on huge amounts of text to predict the next word, which lets it write, summarise, translate and chat.