PH

Phi-4-mini-reasoning

PhiGGUF by MungertUpdated 2025-09-24
Download on Hugging Face
Parameters3.84B
Context128K
Layers32
Architecturephi3
Smallest quant1.2 GB
Quick answer

Phi-4-mini-reasoning needs about 4.1 GB of memory at Q4_K_M (a 2.5 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 30 quants, from 1.2 GB (TQ1_0) to 7.7 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.

Phi-4-mini-reasoning: 3.84B Phi3 model optimized for coding and reasoning

Phi-4-mini-reasoning is a 3.84B parameter Phi3 model. It supports 131072 token context and is released under the MIT license. The model is tuned for coding and reasoning tasks.

Good at

  • Coding tasks
  • Reasoning problems
  • Long‑context handling

Limits

  • No vision support
  • Large VRAM needed for higher precision
  • No mixture of experts

Best for

  • Local code generation
  • Complex problem solving
  • Educational tutoring

Who it’s for. Developers and researchers who need a lightweight Phi model for local inference. The model is ideal for users with moderate GPU memory who want to run reasoning workloads without external services.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
TQ1_01.2 GB 2.8 GB
Fits Get file →
IQ1_S1.2 GB 2.8 GB
Fits Get file →
IQ1_M1.2 GB 2.8 GB
Fits Get file →
IQ2_XXS1.3 GB 2.9 GB
Fits Get file →
TQ2_01.3 GB 2.9 GB
Fits Get file →
IQ2_XS1.4 GB 3.0 GB
Fits Get file →
Q2_K_S1.5 GB 3.1 GB
Fits Get file →
IQ2_S1.6 GB 3.1 GB
Fits Get file →
IQ2_M1.6 GB 3.2 GB
Fits Get file →
IQ3_XXS1.7 GB 3.3 GB
Fits Get file →
IQ3_XS1.8 GB 3.3 GB
Fits Get file →
IQ3_S1.8 GB 3.4 GB
Fits Get file →
Q3_K_S1.8 GB 3.4 GB
Fits Get file →
IQ3_M1.9 GB 3.5 GB
Fits Get file →
Q3_K_M2.1 GB 3.7 GB
Fits Get file →
Q4_02.2 GB 3.7 GB
Fits Get file →
IQ4_XS2.2 GB 3.8 GB
Fits Get file →
IQ4_NL2.3 GB 3.9 GB
Fits Get file →
Q4_K_S2.3 GB 3.9 GB
Fits Get file →
Q4_12.4 GB 4.0 GB
Fits Get file →
Q4_K_M2.5 GB 4.1 GB
Fits Get file →
Q5_02.6 GB 4.2 GB
Fits Get file →
Q5_K_S2.7 GB 4.3 GB
Fits Get file →
Q5_K_M2.8 GB 4.4 GB
Fits Get file →
Q5_12.9 GB 4.5 GB
Fits Get file →
Q6_K_M3.2 GB 4.7 GB
Fits Get file →
Q8_04.1 GB 5.7 GB
Fits Get file →
BF164.7 GB 6.2 GB
Fits Get file →
F164.7 GB 6.2 GB
Fits Get file →
BF167.7 GB 9.3 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_M (file size 2.49 GB, memory 4.1 GB) or Q4_1 (2.41 GB, 4 GB). If 12 GB VRAM is available, Q5_K_M (2.85 GB, 4.4 GB) offers higher precision. For the smallest footprint, IQ1_S (1.2 GB, 2.8 GB) fits 4 GB VRAM but with lower precision.

What GPU do you need for Phi-4-mini-reasoning?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is Q4_1 (2.4 GB file, about 4.0 GB in use).
  • 6 GB VRAM: the best quant that fits is Q8_0 (4.1 GB file, about 5.7 GB in use).
  • 8 GB VRAM: the best quant that fits is BF16 (4.7 GB file, about 6.2 GB in use).
  • 12 GB VRAM: the best quant that fits is BF16 (7.7 GB file, about 9.3 GB in use).

Which Phi-4-mini-reasoning quant should you choose?

Q4_K_M is the usual starting point for Phi-4-mini-reasoning: it balances file size and quality well. Here it is a 2.5 GB file.

The Q8_0 file is 4.1 GB, about 1.6 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is TQ1_0 at 1.2 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturephi3
Parameters3.84B
Layers32
Trained context131,072 tokens
KV cache (fp16)128 KB per token

Source

Quantized byMungert
FamilyPhi
Licensemit
Downloads (HF)1,040
Last updated2025-09-24

Frequently asked questions about Phi-4-mini-reasoning

How much VRAM does Phi-4-mini-reasoning need?

About 4.1 GB at Q4_K_M with an 8K context window. That is the 2.5 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Phi-4-mini-reasoning on an 8 GB GPU?

Yes, with BF16: the 4.7 GB file needs about 6.2 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Phi-4-mini-reasoning GGUF?

It depends on the quant: TQ1_0 is 1.2 GB, IQ1_S is 1.2 GB, IQ1_M is 1.2 GB, IQ2_XXS is 1.3 GB, TQ2_0 is 1.3 GB, IQ2_XS is 1.4 GB and 24 more.

What context length does Phi-4-mini-reasoning support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.

Where can I download Phi-4-mini-reasoning GGUF?

The files are on Hugging Face in the Mungert/Phi-4-mini-reasoning-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Phi-4-mini-reasoning use?

Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Phi-4-mini-reasoning on 8 GB VRAM?

Yes, use Q4_K_M or Q4_1 quantizations. They require approximately 4 GB of VRAM at 8 k context.

Does the model support image input?

No, the model has no vision capabilities.

What license applies to this model?

The model is released under the MIT license.

Similar models

Browse all →

More from Mungert

All by Mungert →

More Phi models

All Phi →

Other small models

All small models →