# What is GPTQ? Definition and meaning

Quantizes weights layer by layer to 4 bits or less, using approximate second-order information to limit error.

## Definition

**GPTQ**: Quantizes weights layer by layer to 4 bits or less, using approximate second-order information to limit error.

In short: Post-training quantization using second-order information.

## When it is used

Compressing language models for cheaper GPU inference.

## Learn more

The full guide to **[GPTQ](/tools/gptq)** explains how it works, when to use it and how it compares with the alternatives, with examples. The [glossary](/glossary) lists every term on the site from A to Z.

Previous term: [GPT-6](/glossary/gpt-6) · Next term: [Gradient boosting](/glossary/gradient-boosting)
