qwen3-8b-4bit
Run Qwen3 8B on a Mac
Run it
Install: pip install rapid-mlx or Rapid-MLX Desktop.
rapid-mlx serve qwen3-8b-4bit
Minimum RAM: No published minimum; 18.5 GB measured peak. Leave macOS and context headroom. Agent setup guide
The previous-generation Qwen3 8B — still one of the most-pulled aliases because a lot of guides reference it. If you're choosing fresh, Qwen3.5 9B is the same footprint and newer.
8 B dense · 4-bitguide compatibility · 16 GB Macs
Alibaba 40K context Base model intelligence 8.3 AA
First request
rapid-mlx serve qwen3-8b-4bit
Weights (4 GB) download on first run; you get an OpenAI-compatible
endpoint at http://localhost:8000/v1. No rapid-mlx yet? It's one line —
curl -fsSL https://rapidmlx.com/install.sh | bash — or take the
desktop app.
Point a client at it. OpenAI SDK: base_url="http://localhost:8000/v1",
model="qwen3-8b-4bit". Claude Code:
ANTHROPIC_BASE_URL=http://localhost:8000 (no /v1 suffix) —
see the Claude Code guide.
Measured on real hardware
Headline machine · Mac mini M4 Pro · 48 GB
All measured machines
| Machine | Decode | First token | Peak memory | Cold boot | Weights |
|---|---|---|---|---|---|
| Mac mini M4 Pro · 48 GB rapid-mlx 0.15.5 · 2026-10-03 raw data | 54.0 tok/s | 5.3 s | 18.5 GB | — | 4 GB |
Each row cites its own run; numbers are never copied across machines. Mac mini M4 Pro · 48 GB: rapid-mlx 0.15.5, 2026-10-03 — raw data + method — 4 concurrent streams: 111.8 tok/s aggregate.
Community runs, one page per Mac: Apple M4 Pro.
Will it fit your Mac?
Measured 18.5 GB Metal peak (4.9 GB process RSS) on a 48 GB M4 Pro across the whole benchmark run — up to a ~16K-token prompt and four concurrent streams, including the engine's prefix cache, which defaults to 20% of RAM. A short chat needs less; long contexts grow the KV cache, so leave headroom.
For the conservative install-default placement see the hardware tiers table; to compare against every model your RAM can hold, use the live picker.
Variants & profiles
qwen3.5-9b-4bit— Qwen3.5 9B — the current generation at the same size
FAQ
How much memory does Qwen3 8B need on a Mac?
Measured 18.5 GB Metal peak (4.9 GB process RSS) on a 48 GB M4 Pro across the whole benchmark run — up to a ~16K-token prompt and four concurrent streams, including the engine's prefix cache, which defaults to 20% of RAM. A short chat needs less; long contexts grow the KV cache, so leave headroom.
How fast is Qwen3 8B on Apple Silicon?
We measured 54.0 tok/s on a Mac mini M4 Pro · 48 GB, rapid-mlx 0.15.5, 2026-10-03; peak memory 18.5 GB on that run. Every row on this page keeps its own chip, version and date — no number is copied across machines.
How do I run Qwen3 8B locally?
rapid-mlx serve qwen3-8b-4bit — the weights download on first run and you get an OpenAI-compatible endpoint at localhost:8000/v1 for Aider, Codex and any OpenAI client, plus the Anthropic API at localhost:8000 for Claude Code.