456
Dictionary

Quantization

Shrinking a model by lowering weight precision.

1 min readupdated 2026-07-04

/ quick answer

Quantization converts 16-bit weights to 8-bit, 4-bit, or lower. It shrinks model size 2-4x with a small accuracy hit — the key trick behind running Llama-scale models on a laptop. Shrinking a model by lowering weight precision.

Shrinking a model by lowering weight precision. Quantization converts 16-bit weights to 8-bit, 4-bit, or lower. It shrinks model size 2-4x with a small accuracy hit — the key trick behind running Llama-scale models on a laptop. In practice: A 70B model quantized to 4-bit fits in ~40GB VRAM and runs on a single H100. This dictionary node is part of the Onexial knowledge graph and links to related concepts, workflows and tools below.
Definition
Quantization converts 16-bit weights to 8-bit, 4-bit, or lower. It shrinks model size 2-4x with a small accuracy hit — the key trick behind running Llama-scale models on a laptop.
Example
A 70B model quantized to 4-bit fits in ~40GB VRAM and runs on a single H100.
/ frequently asked

What is Quantization?

Quantization converts 16-bit weights to 8-bit, 4-bit, or lower. It shrinks model size 2-4x with a small accuracy hit — the key trick behind running Llama-scale models on a laptop.

What is an example of Quantization?

A 70B model quantized to 4-bit fits in ~40GB VRAM and runs on a single H100.

Why does Quantization matter for AI and automation?

Shrinking a model by lowering weight precision. It connects to the workflows, prompts and tool stacks linked on this page, so you can move from definition to execution without leaving Onexial.

/ topics#ai#infra