Run MaziyarPanahi/Phi-4-mini-instruct-GGUF locally
MaziyarPanahi/Phi-4-mini-instruct-GGUF is a mid-size instruction-tuned chat model with 3.84 billion parameters, built on the phi3 architecture. It has been downloaded 164,362 times.
To run MaziyarPanahi/Phi-4-mini-instruct-GGUF locally at a 4,096-token context, its quantized versions need between 2.87 GB (Q2_K, lowest quality) and 8.45 GB (GGUF, highest quality) of memory, weights plus KV cache and a system margin included.
For most users the best balance is Q8_0, needing about 5.1 GB. That means MaziyarPanahi/Phi-4-mini-instruct-GGUF fits entirely in the VRAM of a 6 GB GPU or larger, running fully on the GPU.
All quantizations
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| Q2_K | 3.51 | Fair | 1.57 GB | 0.5 GB | 2.87 GB | 255.3 t/s | Fits in VRAM |
| Q3_K_S | 3.96 | Fair | 1.77 GB | 0.5 GB | 3.07 GB | 226.4 t/s | Fits in VRAM |
| Q3_K_M | 4.42 | Good | 1.97 GB | 0.5 GB | 3.27 GB | 202.8 t/s | Fits in VRAM |
| Q3_K_L | 4.69 | Good | 2.1 GB | 0.5 GB | 3.4 GB | 190.9 t/s | Fits in VRAM |
| Q4_K_S | 4.88 | Good | 2.18 GB | 0.5 GB | 3.48 GB | 183.7 t/s | Fits in VRAM |
| Q4_K_M | 5.2 | Very good | 2.32 GB | 0.5 GB | 3.62 GB | 172.4 t/s | Fits in VRAM |
| Q5_K_S | 5.69 | Very good | 2.54 GB | 0.5 GB | 3.84 GB | 157.5 t/s | Fits in VRAM |
| Q5_K_M | 5.94 | Very good | 2.65 GB | 0.5 GB | 3.95 GB | 150.8 t/s | Fits in VRAM |
| Q6_K | 6.58 | Excellent | 2.94 GB | 0.5 GB | 4.24 GB | 136.1 t/s | Fits in VRAM |
| Q8_0 | 8.52 | Excellent | 3.8 GB | 0.5 GB | 5.1 GB | 105.1 t/s | Fits in VRAM |
| GGUF | 16.02 | Excellent | 7.15 GB | 0.5 GB | 8.45 GB | 7.0 t/s | Offload |
KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.
Frequently asked questions
How much VRAM do you need to run MaziyarPanahi/Phi-4-mini-instruct-GGUF?
You need about 5.1 GB of VRAM to run MaziyarPanahi/Phi-4-mini-instruct-GGUF entirely on the GPU using the Q8_0 quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run MaziyarPanahi/Phi-4-mini-instruct-GGUF on an 8 GB GPU?
Yes. With 8 GB of VRAM you can run MaziyarPanahi/Phi-4-mini-instruct-GGUF fully on the GPU using Q8_0 (about 5.1 GB).
Can I run MaziyarPanahi/Phi-4-mini-instruct-GGUF on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run MaziyarPanahi/Phi-4-mini-instruct-GGUF fully on the GPU using GGUF (about 8.45 GB).
Can I run MaziyarPanahi/Phi-4-mini-instruct-GGUF on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run MaziyarPanahi/Phi-4-mini-instruct-GGUF fully on the GPU using GGUF (about 8.45 GB).
What is the best quantization for MaziyarPanahi/Phi-4-mini-instruct-GGUF?
If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.