Qwen3 8B (Q4_K_M)on CPU Only

Q4_K_M CPU Only

Overview

Qwen3 8B is a 8B parameter dense language model by Qwen, with code, multilingual, thinking, tool-calls capabilities. It supports a context window of up to 40,960 tokens.

Qwen3 8B is an 8-billion-parameter dense transformer from Alibaba's Qwen team, featuring built-in thinking capabilities alongside code generation, tool calling, and multilingual support. It advances beyond Qwen2.5 with improved reasoning, supporting chain-of-thought inference in a compact form factor. The model covers 14 languages including English, Chinese, and Arabic. With a 40K context window and flash attention, it fits on a single consumer GPU and quantizes efficiently for cost-effective self-hosted reasoning workloads.

At Q4_K_M quantization (medium quality tier), the model weighs 4.68 GB. This exceeds the 0 GB of VRAM on CPU Only. Inference is still possible via CPU offload or memory-mapped loading from disk, but expect significantly reduced performance.

A CPU-only configuration with no GPU acceleration. Inference runs entirely on the CPU, which is significantly slower than GPU-accelerated setups but requires no special hardware. Performance and maximum model size depend on available system RAM. Suitable for testing, development, or deployments where no GPU is available.

Hardware Requirements

Model size	4.68 GB
VRAM available	0 GB
VRAM used	0 GB
Min RAM required	4.7 GB
GPU layers	0 / 36
Context size	40,960
Backend	cpu
Flash attention	No

Performance Notes

Reduced performance

36 of 36 layers run on CPU. Your system needs at least 4.7 GB RAM. If system RAM is insufficient, data will be read from disk on demand, with significant performance impact.

Deploy

Server Local

Command

helmfile --state-values-file <(curl -s https://www.prositronic.eu/values/qwen3-8b/q4_k_m/cpu.yaml) apply

Generated values.yaml

/values/qwen3-8b/q4_k_m/cpu.yaml

Loading values…

Install llama.cpp

Install llama.cpp from the official build instructions for your platform and backend.

Download Model

curl -L -o qwen3-8b.gguf "https://huggingface.co/unsloth/Qwen3-8B-GGUF/resolve/main/Qwen3-8B-Q4_K_M.gguf"

Start Server

llama-server \
  -m qwen3-8b.gguf \
  --n-gpu-layers 0 \
  --ctx-size 40960

Verify

curl http://localhost:8080/health

Explore

• Configure a deployment
• Browse models
• Browse hardware

Other quantizations

• Qwen3 8B (Q8_0)
• Qwen3 8B (Q5_K_M)

Other hardware

• Qwen3 8B on NVIDIA A100 80GB
• Qwen3 8B on NVIDIA H100
• Qwen3 8B on NVIDIA RTX 3090

Frequently Asked Questions

How much VRAM does Qwen3 8B (Q4_K_M) need?

The Q4_K_M quantization of Qwen3 8B requires 4.68 GB. The 0 GB of VRAM on CPU Only is insufficient for GPU layers, so inference runs on CPU.

Can I run Qwen3 8B on CPU Only?

It is possible but not recommended. CPU Only does not have enough VRAM to accelerate Qwen3 8B (Q4_K_M), so inference will rely on CPU and system RAM.

What is quantization?

Quantization reduces a model's numerical precision from its original floating-point format to a more compact representation. This shrinks the file size and VRAM footprint, making it possible to run large models on consumer hardware. The trade-off is a small reduction in output quality. Q4_K_M compresses Qwen3 8B from its original size down to 4.68 GB.

What quantization should I choose for Qwen3 8B?

Q4_K_M is a medium-quality quantization. Higher-quality quants (Q8, Q6) preserve more model accuracy but need more VRAM. Lower quants (Q4, Q3, Q2) reduce VRAM usage at the cost of some quality. Choose based on your available hardware and quality requirements.

Why are some layers offloaded to CPU?

CPU Only has 0 GB of VRAM, but Qwen3 8B (Q4_K_M) requires approximately 4.68 GB. Only 0 of 36 layers fit in VRAM; the remaining layers run on CPU, which is slower but still functional.

How do I run Qwen3 8B (Q4_K_M) with Ollama?

Run ollama run qwen3:8b-q4_k_m to start Qwen3 8B (Q4_K_M). Ollama handles downloading the model weights automatically on first run.

Last updated: March 5, 2026