Phi-4 mini 3.8B VRAM requirements
Microsoft · 3.8B params · MIT. Estimates, not vendor numbers.
Phi-4 mini 3.8B (3.8B parameters) needs about 3.8 GB of VRAM or unified memory at Q4, 4.2 GB at Q5 and 5.5 GB at Q8 (weights plus about 1.5 GB headroom, our estimate). The smallest common setup that runs it is 8GB GPU (RTX 4060).
| Quant | Download | Memory needed |
|---|---|---|
| Q4 | 2.3 GB | 3.8 GB |
| Q5 | 2.7 GB | 4.2 GB |
| Q8 | 4.0 GB | 5.5 GB |
Will it run on my machine?
| Setup | Runs? | Speed est. |
|---|---|---|
| 8GB GPU (RTX 4060) | Yes, Q8 | ~40 tok/s |
| 12GB GPU (RTX 3060 / 4070) | Yes, Q8 | ~75 tok/s |
| 16GB GPU (RTX 4060 Ti 16GB) | Yes, Q8 | ~43 tok/s |
| 24GB GPU (RTX 4090) | Yes, Q8 | ~150 tok/s |
| 32GB GPU (RTX 5090) | Yes, Q8 | ~266 tok/s |
| Mac 16GB (M4) | Yes, Q8 | ~18 tok/s |
| Mac 48GB (M4 Pro) | Yes, Q8 | ~41 tok/s |
| Mac 128GB (M4 Max) | Yes, Q8 | ~81 tok/s |
| PC, 32GB RAM, no GPU | Yes, Q5 | ~13 tok/s |
Best quant that fits each setup, with estimated generation speed from memory bandwidth. Real speed depends on the runtime, context length and settings.
Similar-size models
Ministral 3 3B (3B) · Qwen3 4B (4B) · Gemma 4 E4B (8B) · Ministral 3 8B (8B) · Llama 3.1 8B Instruct (8B) · Qwen3 8B (8.2B)
FAQ
How much VRAM does Phi-4 mini 3.8B need?
About 3.8 GB at Q4, 4.2 GB at Q5 and 5.5 GB at Q8, including about 1.5 GB headroom. These are estimates from the parameter count on the official model card.
Can Phi-4 mini 3.8B run on 8GB of VRAM?
Yes, at Q8.
Can Phi-4 mini 3.8B run on a Mac?
Yes. Mac 16GB (M4), Mac 48GB (M4 Pro), Mac 128GB (M4 Max) can run it.
Which quant of Phi-4 mini 3.8B should I use?
Use the highest quant that fits with room to spare: Q8 is near full quality, Q5 is a good balance, and Q4 is the smallest that usually keeps quality acceptable.