Skip to content
localmodel.run

Text model · Qwen2.5

Qwen2.5 0.5B requirements

Qwen2.5 family · 0.494B params · released 2024-09-19 · 23.2M Ollama pulls. Minimum to run at Q4_K_M: Apple M1 (8GB).

LicenseApache-2.0· Commercial OK↓ 2.9M/mo♥ 531on HuggingFace
Q4_K_M
0.491 GB
Q8_0
0.676 GB
Total @ Q4 (4k)
~1.5 GB
Context
128 k

Quantization sizes

GGUF quantson disk
QuantizationSize on disk
Q2_K0.2 GB est
Q3_K_M0.2 GB est
Q4_K_M (default)0.491 GB
Q5_K_M0.4 GB est
Q6_K0.4 GB est
Q8_00.676 GB
FP161 GB est

Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.

Run it

Ollama
$ ollama run qwen2.5:0.5b
llama.cpp
$ llama-cli -hf Qwen/Qwen2.5-0.5B-Instruct-GGUF:Q4_K_M
LM Studio
$ lms get Qwen/Qwen2.5-0.5B-Instruct-GGUF

Which devices can run Qwen2.5 0.5B?

FAQ

How much VRAM or RAM does Qwen2.5 0.5B need?

At Q4_K_M, Qwen2.5 0.5B needs about 1.5 GB (weights ~0.491 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~1.7 GB.

Can Qwen2.5 0.5B run on a laptop?

Yes, Qwen2.5 0.5B fits on a 16 GB machine at Q4_K_M and runs on Apple Silicon or a 12 GB+ GPU comfortably.

Can I use Qwen2.5 0.5B commercially?

Yes. Qwen2.5 0.5B is licensed Apache-2.0, which permits commercial use.

Smallest Qwen2.5 dense model. Trained on 18T tokens, supports 29+ languages, 128K context (generates up to 8K). Ideal for on-device and microcontroller-class edge inference.

Sources

Memory figures are estimates. See methodology.