Budget build · mATX tower · New parts
GPUter
Budget single-GPU Blackwell inference and training rig
Built by Llama Builds · Updated
- Total cost
- $1,250
- VRAM
- 16GB
- GPUs
- 1
- System RAM
- 32GB
What can this build run?
With 16GB of VRAM across one GPU, this build comfortably runs models up to Phi-4 (Q5_K_M, ~33 tokens/sec). Our top all-round pick for it is Qwen3.8 27B at Q3_K_M (~26 tokens/sec).
Parts list
Every part in the GPUter
We may earn a commission when you buy through links on this page. It never changes what we recommend. How we make money
| Part | Price | Buy |
|---|---|---|
GPU NVIDIA GeForce RTX 5060 Ti 16GB | $450 | Amazon |
Motherboard MSI PRO B550M-VC WiFi ProSeries | $140 | Amazon |
CPU AMD Ryzen 5 5500 | $80 | Amazon |
Memory CORSAIR Vengeance LPX DDR4 32GB | $90 | Amazon |
Storage M.2 SSD 4TB | $280 | Amazon |
Case JONSBO/JONSPLUS Z20 mATX | $110 | Amazon |
Power Supply CORSAIR RM750e | $100 | Amazon |
| Total | $1,250 |
Prices at time of listing (2025); check current prices. Our own build cost $1,040.
Capability
Models this build runs well
Estimated for a single user with 8K context. Best quantization that fits, ranked by our capability score.
Best overall · also coding, reasoning, vision, largest that fits
Qwen3.8 27BQ3_K_M · ~26 tokens/sec
Runner-up
Qwen3.5 9BQ8_0 · ~32 tokens/sec
Fastest good model
Gemma 4 26B A4BQ3_K_M · ~90 tokens/sec
| Model | Quant | Speed | Fit |
|---|---|---|---|
| Qwen3.8 27B | Q3_K_M | ~26 tok/s | Fits, tight |
| Qwen3.5 9B | Q8_0 | ~32 tok/s | Runs well |
| gpt-oss-20b | MXFP4 | ~74 tok/s | Fits, tight |
| Gemma 4 26B A4B | Q3_K_M | ~90 tok/s | Fits, tight |
| Gemma 4 E4B | Q8_0 | ~36 tok/s | Runs well |
| Phi-4 | Q5_K_M | ~33 tok/s | Runs well |
The build
About this build
Our answer to the most affordable useful AI computer in 2025, with an RTX 5060 Ti 16GB, a Ryzen 5 5500, and 32GB of RAM in a compact JONSBO case. It is whisper-quiet and built entirely from new parts.
Why this build
What’s the most affordable AI computer with a GPU (a “GPUter”) you can build in 2025? We have an answer. That’s our actual GPUter in the photo, built for just $1,040 with all new parts.
It’s whisper-quiet but packs serious power. It’s a great entry-level build for AI enthusiasts who want capable performance without breaking the bank, and the 16GB RTX 5060 Ti makes it a budget Blackwell rig for both inference and training.
Notes and caveats
- We paid $1,040 for our build. The parts list totals $1,250 at the listed prices, so check current pricing before you buy.
Specifications
Full specs
- GPU
- RTX 5060 Ti 16GB
- CPU
- AMD Ryzen 5 5500
- Motherboard
- MSI PRO B550M-VC WiFi
- Memory
- 32GB DDR4 (Corsair Vengeance LPX)
- Storage
- 4TB M.2 SSD
- Power supply
- Corsair RM750e
- Case
- JONSBO Z20 (mATX)
- Cooling
- AMD stock cooler
- Form factor
- mATX tower
- GPU power
- 180W rated
Keep browsing
