BTC
ETH
SOL
XRP
BNB
TON
DOT
LTC
POL
HBAR
NEAR
ARB
UNI
Shrinking an AI model by storing its numbers with less precision, so it runs faster and on smaller hardware with a small loss of quality.
AI Drafted with AI help · Reviewed by Keshav Aggarwal · Updated
In simple words
Imagine writing 3.14159265 as just 3.14. It takes less space and is close enough for most sums.
Quantisation does that to the billions of numbers inside an AI model.
The technical side
Model weights are usually trained in 16- or 32-bit formats. Quantisation stores them in 8, 4 or even fewer bits, cutting memory needs by half or more and speeding up inference. Formats such as GGUF let quantised open-weights models run on laptops and phones.
Heavier quantisation saves more but can hurt accuracy, especially on maths and long reasoning.
Quantisation kya hai?
Quantisation model ke numbers ko chhota karke model ko halka banata hai, taaki laptop pe bhi chale. Thodi quality kam ho sakti hai.
Related terms
- ParametersAIThe numbers inside an AI model that it learned in training. Model size is counted in parameters, often billions of them.
- Open WeightsAIAn AI model whose trained numbers (weights) are published, so anyone can download it and run it on their own computers.
- InferenceAIThe moment an AI model is used to answer: it takes your input and works out the output. Each answer you get is one inference.