
RTX 4060 8 GB
Budget8 GB VRAM · ~7.2 GB for a model
NVIDIA GeForce RTX 4060 8 GB
- Qwen2.5-1.5B-Instruct~164 tok/s
- Llama-3.2-1B-Instruct~185 tok/s
- gemma-3-1b-it~204 tok/s
The one spec that decides which AI models you can run at home is memory: VRAM on a graphics card, or unified memory on a Mac. Each card shows the most downloaded open models that fit and their estimated speed.
As an Amazon Associate we earn from qualifying purchases. Prices change often, so we link to Amazon instead of listing them.
For a desktop PC. VRAM decides which models fit; more of it runs bigger models.

8 GB VRAM · ~7.2 GB for a model
NVIDIA GeForce RTX 4060 8 GB

12 GB VRAM · ~10.8 GB for a model
NVIDIA GeForce RTX 3060 12 GB
Apple silicon shares one pool of unified memory between the chip and the GPU, so a model can use most of it.

16 GB unified memory · ~11.2 GB for a model
M5, 10-core GPU

16 GB unified memory · ~11.2 GB for a model
M4

24 GB unified memory · ~16.8 GB for a model
M4
| Device | Type | Memory | Usable | Budget | Shop |
|---|---|---|---|---|---|
| RTX 4060 8 GB | Graphics cards | 8 GB VRAM | ~7.2 GB | Budget | Check price |
| RTX 3060 12 GB | Graphics cards | 12 GB VRAM | ~10.8 GB | Budget | Check price |
| MacBook Air with 16 GB | Macs | 16 GB unified memory | ~11.2 GB | Budget | Check price |
| Mac mini with 16 GB | Macs | 16 GB unified memory | ~11.2 GB | Budget | Check price |
| Mac mini with 24 GB | Macs | 24 GB unified memory | ~16.8 GB | Budget | Check price |
Usable memory is about 70% of a Mac's unified memory and 90% of a graphics card's VRAM. Models are the most downloaded open language models on Hugging Face that fit. Speed data: ModelFit (CC BY 4.0), modelled decode estimates for a 4-bit build at a 4K context, not measured benchmarks and not an endorsement by ModelFit.
As an Amazon Associate we earn from qualifying purchases.
Photo credits (Wikimedia Commons, resized): GeForce RTX 4060 graphics card: 极客湾Geekerwan, CC BY 3.0