AMD GPUs
AMD Radeon RX 7900 XTX
- Memory
- 24 GB
- Bandwidth
- 960 GB/s
Largest popular model at 4-bit: Qwen3.6 27B
24 GB and 960 GB/s: RTX 3090-class memory, bought new.
from $1,358 (typical price)Recommended
Apple · Apple silicon · 2026
Apple Mac mini (M6)
Illustration
Good budget option
A silent, tiny always-on box for small models. Choose 32 GB if AI is the goal.
- Memory
- 16 GB–32 GB
- Bandwidth
- 170 GB/s
- Power
- —
- Largest popular model at 4-bit
- Qwen3.6 27B
from $899 (launch price)
Researched from manufacturer data and public sources; not tested by us.· Last reviewed Oct 11, 2026
What it can run
EstimatePopular models at Q4_K_M with a 32K-token context. Memory needed from real model files; speed is the theoretical ceiling.
Change model size or context
Model weightsContext cache (KV)RuntimeUsable limit (colour = verdict)
| Model | Verdict | Memory | Speed ceiling |
|---|---|---|---|
| Qwen3 1.7B | Runs | ≤ 153 tok/s | |
| Qwen3.5 9B | Runs | ≤ 30 tok/s | |
| Gemma 4 12B | Runs | ≤ 24 tok/s | |
| gpt-oss-20b | Runs | ≤ 85 tok/s | |
| Qwen3.6 27B | Runs, tight | ≤ 10 tok/s | |
| Qwen3.6 35B A3B | Not enough memory | — | |
| Llama 3.3 70B | Not enough memory | — | |
| gpt-oss-120b | Not enough memory | — | |
| Qwen3.5 122B A10B | Not enough memory | — | |
| Qwen3.5 397B A17B | Not enough memory | — |
Theoretical maximum for generating text: memory bandwidth divided by the bytes read per token. Real runs usually reach about half to three quarters of it.
Where to buy
Prices are editorial references (launch price and a typical street price on the date shown), not live prices. In 2026 the memory shortage has pushed many prices well above launch.
- Mac mini M6, 16 GB16 GBLaunch price: $899
- Check price Amazon.com (opens in a new tab)Amazon.comCheck availability
Other regions (3)
- Check price Amazon.co.uk (opens in a new tab)Amazon.co.ukUnited Kingdom · Check availability · Search results
- Check price Amazon.es (opens in a new tab)Amazon.esSpain · Check availability
- Check price Amazon.de (opens in a new tab)Amazon.deGermany · Check availability · Search results
- Mac mini M6, 32 GB32 GB
- Check price Amazon.com (opens in a new tab)Amazon.comCheck availability · Search results
Other regions (3)
- Check price Amazon.co.uk (opens in a new tab)Amazon.co.ukUnited Kingdom · Check availability · Search results
- Check price Amazon.es (opens in a new tab)Amazon.esSpain · Check availability · Search results
- Check price Amazon.de (opens in a new tab)Amazon.deGermany · Check availability · Search results
Prices are shown only if updated in the last 24 hours and may change; the retailer’s price at checkout applies. We may earn a commission from these links at no cost to you. Affiliate disclosure
Overview
Editorial opinionBest for: Home Assistant voice, small private chat models and a low-power home server.
Editorial opinionWould we buy it today?
Yes for small models and low power; for 27B+ models get an M5 Pro or more.
Pros and cons
Editorial opinionPros
- Near-silent and a few watts at idle
- Simple setup with Ollama or LM Studio
- MLX support
Cons
- 170 GB/s: slow with anything over 14B
- Memory cannot be upgraded later
Good to know
macOS lets the GPU use about two thirds of memory on machines with up to 36 GB, and about three quarters above that. The limit can be raised (sudo sysctl iogpu.wired_limit_mb=…), leaving less for macOS and other apps; our figures use the default.
Specifications
Factual data- Manufacturer
- Apple
- Chip
- M6
- Architecture
- Apple silicon
- Type
- Unified-memory computer
- Form factor
- Mini PC
- Released
- 2026
- Availability
- On sale
- Memory options
- 16 GB / 24 GB / 32 GB
- Memory type
- Unified LPDDR
- Memory bandwidth
- 170 GB/s
- Usable for models (default)
- 21 GB / 32 GB
- Compute stack
- Metal / MLX
- Models
- Any model that fits
- Several units for one model
- Yes
Source: Manufacturer specifications
Software support
Editorial opinionHow well the popular local-AI tools run on it today (editorial assessment).
- llama.cppExcellent
- OllamaExcellent
- LM StudioExcellent
- vLLMPoor
- MLXExcellent
Measured speeds
Factual dataMeasured results with their source. We never invent speeds.
No measured results yet. The table above shows the theoretical ceiling instead.
Alternatives
Intel GPUs
Intel Arc Pro B70
- Memory
- 32 GB
- Bandwidth
- 608 GB/s
Largest popular model at 4-bit: Qwen3.6 35B A3B
Intel's big Battlemage: 32 GB and 608 GB/s for less than AMD's R9700.
from $1,268 (typical price)Recommended
NVIDIA GPUs
NVIDIA GeForce RTX 3090
- Memory
- 24 GB
- Bandwidth
- 936 GB/s
Largest popular model at 4-bit: Qwen3.6 27B
The homelab classic: 24 GB and 936 GB/s, bought used. Two of them make 48 GB.
from $1,340 (typical price)Good budget option