Skip to content

Qwen2.5-Coder-1.5B-Instruct-GPTQ-Int4

A 1.5B code model with GPTQ 4-bit quantization

Visit official site

What Qwen2.5-Coder-1.5B-Instruct-GPTQ-Int4 is

The model is described as a Qwen2.5-Coder series release built for code generation, code reasoning and code repair, based on Qwen2.5.

What you can do with it

  • Write code for a requested function
  • Explain how existing code works

Getting started

  • Install a compatible Hugging Face transformers version
  • Load weights and tokenizer with the provided classes

Pros & cons

✓ What we like

  • Supports a broad range of mainstream programming languages
  • Uses GPTQ 4-bit quantization for size and speed

! What to watch out for

  • The Hugging Face transformers version must meet requirements
  • Loading needs both AutoModelForCausalLM and AutoTokenizer

FAQ

What quantization is used?

The material names GPTQ 4-bit quantization.

How large is the model?

It is listed at 1.54B parameters with 28 layers.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review