Best laptops for local LLMs in the UK
For local LLMs a laptop's GPU memory matters most, and laptop GPUs top out at 24GB (RTX 5090 Laptop). Below are the UK laptops at Scan and Box with the most VRAM first, then cheapest, checked 2026-10-16.
| Laptop | GPU | TGP | Biggest model (est.) | Price |
|---|---|---|---|---|
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £2,799.99 Box, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,099.00 Laptops Direct, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,129.99 Box, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,199.00 Laptops Direct, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,199.99 Scan, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,249.97 Laptops Direct, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,299.98 Laptops Direct, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,299.98 Laptops Direct, 2026-10-16 | |
| RTX 5090 24 GB | 175 W | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,299.99 Scan, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,299.99 Scan, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,299.99 Scan, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,495.00 Box, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,499.99 Scan, 2026-10-16 | |
| RTX 5090 24 GB | 175 W | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,549.98 Scan, 2026-10-16 | |
| RTX 5090 24 GB | unknown | Q4: Qwen2.5-Coder 32B Q8: Qwen3 14B | £3,599.97 Laptops Direct, 2026-10-16 |
FAQ
How much VRAM do I need in a laptop for local LLMs?
8GB runs 7–8B models at Q4; 12GB fits 14B at Q4; 16GB fits 14B at Q8 or ~24B at Q4; 24GB (RTX 5090 Laptop) fits ~32B at Q4. These are our estimates using real GGUF file sizes.
Is a MacBook better for local LLMs?
Macs share system memory with the GPU, so high-memory models can load bigger LLMs, though roughly 70% of memory is usable for the GPU by default and speed differs. We don't track Mac laptop prices yet.
Data: Scan UK listings (checked 2026-10-16), NVIDIA/TechPowerUp specs, Hugging Face GGUF sizes. Model fit is an estimate. All GPU deals · All laptops · All models