Can I Run Phi-4 on a NVIDIA RTX 3080 10GB?
Yes
Runs at Q4_K_M — good quality with reasonable headroom.
Model size
14.7B
GPU memory
10.0GB
Smallest quant
Q2_K
Best fit
Q4_K_M
7 quantizations fit your 10.0GB
| Quant | Min VRAM | Recommended | File size | Headroom |
|---|---|---|---|---|
| Q4_K_MBEST | 9.9 GB | 11.4 GB | 9.1 GB | +0.1 GB |
| Q4_K_S | 9.4 GB | 10.9 GB | 8.4 GB | +0.6 GB |
| Q4_0 | 9.3 GB | 10.8 GB | 8.4 GB | +0.7 GB |
| Q3_K_L | 7.5 GB | 9.0 GB | 7.9 GB | +2.5 GB |
| Q3_K_M | 7.2 GB | 8.7 GB | 7.4 GB | +2.8 GB |
| Q3_K_S | 6.7 GB | 8.2 GB | 6.5 GB | +3.3 GB |
| Q2_K | 5.8 GB | 7.3 GB | 5.5 GB | +4.2 GB |
Try it in the cloud first
Don't want to download Phi-4 just to try it? Use a hosted API or rent a GPU by the second.
Affiliate links — we earn a commission at no cost to you.
Advertisement
Full model details
Phi-4 →
All quant variants, benchmark scores, and use-case tags.
Best models for this GPU
NVIDIA RTX 3080 10GB →
Top-ranked open-source models that fit in 10.0GB.
FAQ
Can the NVIDIA RTX 3080 10GB run Phi-4?
Yes. The NVIDIA RTX 3080 10GB's 10.0GB of VRAM is enough to run Phi-4 at Q4_K_M quantization (9.9GB required).
What's the best quantization to use?
Q4_K_M is the highest-precision quantization that fits in your 10.0GB. It uses about 9.9GB of memory and 11.4GB recommended for comfortable inference.
What if I need more headroom for context length?
KV cache memory grows with context length. The numbers above assume a baseline 2K-4K context. For long-context use (32K+), add another 2-6GB depending on the model architecture.