PH

microsoft.Phi 4 Reasoning Vision 15B

PhiGGUF by DevQuasarUpdated 2026-03-08
Download on Hugging Face
Parameters14.7B
Context32K
Layers40
Architecturephi3
Smallest quant5.5 GB
Quick answer

microsoft.Phi 4 Reasoning Vision 15B needs about 11.2 GB of memory at F16 (a 9.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 6 quants, from 5.5 GB (F16) to 15.6 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 6.7 GB.

Phi-4-Reasoning-Vision 15B: multimodal reasoning with text and images

Phi-4-Reasoning-Vision 15B is a 14.7 B parameter model built on the phi3 architecture. It combines a SigLIP‑2 vision encoder with a Phi‑4‑Reasoning backbone using mid‑fusion. The model supports 32,768‑token context and is trained on a curated mix of reasoning and vision data.

Good at

  • Reasoning with chain‑of‑thought
  • High‑resolution image understanding
  • Multimodal text‑image tasks

Limits

  • Requires large VRAM (≈7.7 GB at 8 k context for F16)
  • No audio or video input
  • MIT license limits commercial use without attribution

Best for

  • Fine‑grained document analysis
  • GUI grounding and object detection
  • Captioning and visual question answering

Who it’s for. Developers and researchers needing a compact multimodal model that can perform both reasoning and vision tasks within a single inference pipeline. The model is suitable for applications that require high‑resolution image understanding and chain‑of‑thought reasoning.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
F165.5 GB 7.7 GB
Tight Get file →
F167.4 GB 9.5 GB
Too big Get file →
F169.1 GB 11.2 GB
Too big Get file →
F1610.6 GB 12.8 GB
Too big Get file →
F1612.0 GB 14.2 GB
Too big Get file →
F1615.6 GB 17.8 GB
Too big Get file →

For 8 GB VRAM use the F16 quant with file size 5.55 GB (≈7.7 GB memory at 8 k context). For 12 GB VRAM use the 10.6 GB file (≈12.8 GB memory). Larger VRAM allows the 15.58 GB file (≈17.8 GB memory).

What GPU do you need for microsoft.Phi 4 Reasoning Vision 15B?

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: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: the best quant that fits is F16 (5.5 GB file, about 7.7 GB in use).
  • 12 GB VRAM: the best quant that fits is F16 (9.1 GB file, about 11.2 GB in use).
  • 16 GB VRAM: the best quant that fits is F16 (12.0 GB file, about 14.2 GB in use).
  • 24 GB VRAM: the best quant that fits is F16 (15.6 GB file, about 17.8 GB in use).

Which microsoft.Phi 4 Reasoning Vision 15B quant should you choose?

F16 (9.1 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

If memory is tight, the smallest file is F16 at 5.5 GB, with the quality trade-offs that come with fewer bits. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturephi3
Parameters14.7B
Layers40
Trained context32,768 tokens
KV cache (fp16)200 KB per token

Source

Quantized byDevQuasar
FamilyPhi
Licensemit
Downloads (HF)1,237
Last updated2026-03-08

Frequently asked questions about microsoft.Phi 4 Reasoning Vision 15B

How much VRAM does microsoft.Phi 4 Reasoning Vision 15B need?

About 11.2 GB at F16 with an 8K context window. That is the 9.1 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 microsoft.Phi 4 Reasoning Vision 15B on an 8 GB GPU?

Yes, with F16: the 5.5 GB file needs about 7.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of microsoft.Phi 4 Reasoning Vision 15B GGUF?

It depends on the quant: F16 is 5.5 GB, F16 is 7.4 GB, F16 is 9.1 GB, F16 is 10.6 GB, F16 is 12.0 GB, F16 is 15.6 GB.

What context length does microsoft.Phi 4 Reasoning Vision 15B support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 1.7 GB at 8K and 6.7 GB at 32K.

Where can I download microsoft.Phi 4 Reasoning Vision 15B GGUF?

The files are on Hugging Face in the DevQuasar/microsoft.Phi-4-reasoning-vision-15B-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 microsoft.Phi 4 Reasoning Vision 15B 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‑Reasoning‑Vision on 8 GB VRAM?

Yes, the 5.55 GB F16 quant uses about 7.7 GB memory at 8 k context, fitting within 8 GB VRAM.

Does the model support video input?

No, the model only accepts text and image inputs.

Is the model licensed for commercial use?

It is MIT licensed; commercial use is allowed with attribution.

Other GGUFs of Phi-4-reasoning-vision-15B

Similar models

Browse all →

More from DevQuasar

All by DevQuasar →

More Phi models

All Phi →

Other medium models

All medium models →