Only with CPU offload

Qwen3 Coder 30B-A3B at Q4_K_M needs 18.5 GB but only 9.7 GB is addressable, so about 51% of the layers would stream from system RAM at roughly 60 GB/s. Expect around 13 tokens per second — usable for batch work, painful for chat.

49% on GPU 8K context Q4_K_M · 17.1 GB Apache 2.0 Released Jul 2025

Agentic coding MoE with a 256K native window. Still the most-downloaded local code model.

What hardware do I need for Qwen3 Coder 30B-A3B? →

Fits instead: Qwen3 8B (6.3 GB) · Qwen3 4B (4.0 GB)

The VRAM budget

weights 17.1 GB

Weights 17.1 GB KV cache @ 8K 0.75 GB Runtime overhead 0.6 GB Over budget 8.8 GB past 9.7 GB

Quantisation ladder

Quant	Weights	Total @ 8K	Max context	Tok/s	Quality	Fit
Q8_0	30.2 GB	31.5 GB	—	~5.6	−0.1% ppl	21.8 GB over
Q6_K	23.3 GB	24.6 GB	—	~8.1	−0.4% ppl	14.9 GB over
Q5_K_M	20.1 GB	21.5 GB	—	~10	−0.8% ppl	11.8 GB over
Q4_K_M	17.1 GB	18.5 GB	—	~13	−1.9% ppl	8.8 GB over
Q3_K_M	13.9 GB	15.2 GB	—	~20	−5.4% ppl	5.5 GB over
Q2_K	11.9 GB	13.2 GB	—	~28	−15% ppl	3.5 GB over

Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.

How to run it

Ollama llama.cpp LM Studio vLLM

terminal

$ llama-server \
    -hf Qwen/Qwen3-Coder-30B-A3B-Instruct:Q4_K_M \
    -c 8192 -ngl 23

The engine underneath most of the others. Every knob is exposed. More on llama.cpp.

01Download is 17.1 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.

02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.

03If it falls back to CPU silently, drop the context first, then step down a quantisation.

See all models for this rig Compare with another model