Outmake Models
Models

Phi 3 Medium 128K

Microsoft · May 2024

Get Hugging Face

A 14B chat model from Microsoft.

Downloads
12K
Likes
66
Parameters
14B
Context
128K
Licence
mit

Runs well on M4 · 16 GB

The Q4_K_L size from bartowski is the one to get: a 8.7 GB download that uses 9.8 GB of memory, and writes about 4.5 words a second.

Ratings

KnowledgeMMLU-Pro
51.9%

MMLU-Pro: graduate-level questions in 14 subjects. Scores are the ones filed on the model's Hugging Face page or the board's own table, read 2026-11-05; some come from the maker.

Phi 3 Medium 128K runs well from a Mac with an M1 and 8 GB of memory. 59 of 59 Mac configurations run it well.

Compare with

The nearest models in size.

Sizes

From mmnga/Phi-3-medium-128k-instruct-gguf on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
IQ1_S3.0 GB4.0 GBestimatedOpen
IQ1_M3.2 GB4.3 GBestimatedOpen
IQ2_XXS3.7 GB4.8 GBestimatedOpen
IQ2_XS4.1 GB5.2 GBestimatedOpen
IQ2_S4.3 GB5.4 GBestimatedOpen
IQ2_M4.7 GB5.8 GBestimatedOpen
Q2_K5.1 GB6.2 GBestimatedOpen
IQ3_XXS5.5 GB6.5 GBestimatedOpen
IQ3_XS5.8 GB6.9 GBestimatedOpen
Q3_K_S6.1 GB7.1 GBestimatedOpen
IQ3_S6.1 GB7.1 GBestimatedOpen
IQ3_M6.5 GB7.6 GBestimatedOpen
Q3_K_M6.9 GB8.0 GBestimatedOpen
IQ4_XS7.5 GB8.5 GBestimatedOpen
Q3_K_L7.5 GB8.6 GBestimatedOpen
Q4_07.9 GB9.0 GBestimatedOpen
IQ4_NL7.9 GB9.0 GBestimatedOpen
Q4_K_S8.0 GB9.0 GBestimatedOpen
Q4_K_M8.6 GB9.6 GBestimatedOpen

Run It Anywhere

Phi 3 Medium 128K is an open model, so any app that runs GGUF files can run it. These take the Q4_K_L size from bartowski straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/bartowski/Phi-3-medium-128k-instruct-GGUF:Q4_K_L
  • llama.cppMac, Windows, Linuxllama-server -hf bartowski/Phi-3-medium-128k-instruct-GGUF:Q4_K_L
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/bartowski/Phi-3-medium-128k-instruct-GGUF:Q4_K_L
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBNone worth runningneeds 6.5 GB
8 GBIQ3_M6.5 GB7.6 GB
10 GBQ4_K_M8.6 GB9.6 GB
11 GBQ4_K_M8.6 GB9.6 GB
12 GBQ4_K_M8.6 GB9.6 GB
16 GBQ4_K_M8.6 GB9.6 GB
20 GBQ4_K_M8.6 GB9.6 GB
24 GBQ4_K_M8.6 GB9.6 GB
32 GBQ4_K_M8.6 GB9.6 GB
48 GBQ4_K_M8.6 GB9.6 GB

The largest size, of those that keep three bits a weight or more, whose memory at an 8K chat is under 92% of the card's; the rest is the desktop's. Estimated from the files, the same way as for a Mac; speed on a PC is not estimated yet. A size that doesn't fit the card still runs, slower, with part of it in ordinary memory.

On every Mac
MacSizeFitMemoryWrites
M6 · 16 GBQ4_K_LRuns well9.8 GB6.3
M6 · 24 GBQ4_K_LRuns great11 GB6.3
M6 · 32 GBQ4_K_LRuns great13 GB6.3
M5 · 16 GBQ4_K_LRuns well9.8 GB5.7
M5 · 24 GBQ4_K_LRuns great11 GB5.7
M5 · 32 GBQ4_K_LRuns great13 GB5.7
M5 Pro · 24 GBQ4_K_LRuns great11 GB11
M5 Pro · 48 GBQ4_K_LRuns great17 GB11
M5 Pro · 64 GBQ4_K_LRuns great24 GB11
M5 Max · 36 GBQ4_K_LRuns great17 GB17
M5 Max · 48 GBQ4_K_LRuns great17 GB22
M5 Max · 64 GBQ4_K_LRuns great24 GB22
M5 Max · 128 GBQ4_K_LRuns great24 GB22
M5 Ultra · 96 GBQ4_K_LRuns great24 GB42
M5 Ultra · 256 GBQ4_K_LRuns great24 GB42
M5 Ultra · 512 GBQ4_K_LRuns great24 GB42
M4 · 16 GBQ4_K_LRuns well9.8 GB4.5
M4 · 24 GBQ4_K_LRuns great11 GB4.5
M4 · 32 GBQ4_K_LRuns great13 GB4.5
M4 Pro · 24 GBQ4_K_LRuns great11 GB9.7
M4 Pro · 48 GBQ4_K_LRuns great17 GB9.7
M4 Pro · 64 GBQ4_K_LRuns great24 GB9.7
M4 Max · 36 GBQ4_K_LRuns great17 GB14
M4 Max · 48 GBQ4_K_LRuns great17 GB18
M4 Max · 64 GBQ4_K_LRuns great24 GB18
M4 Max · 96 GBQ4_K_LRuns great24 GB14
M4 Max · 128 GBQ4_K_LRuns great24 GB18
M3 · 8 GBIQ2_XSRuns well5.2 GB5.3
M3 · 16 GBQ4_K_LRuns well9.8 GB3.8
M3 · 24 GBQ4_K_LRuns great11 GB3.8
M3 Pro · 18 GBQ3_K_LRuns great8.6 GB6
M3 Pro · 36 GBQ4_K_LRuns great17 GB5.6
M3 Max · 36 GBQ4_K_LRuns great17 GB11
M3 Max · 48 GBQ4_K_LRuns great17 GB14
M3 Max · 64 GBQ4_K_LRuns great24 GB14
M3 Max · 96 GBQ4_K_LRuns great24 GB11
M3 Max · 128 GBQ4_K_LRuns great24 GB14
M3 Ultra · 96 GBQ4_K_LRuns great24 GB25
M3 Ultra · 256 GBQ4_K_LRuns great24 GB25
M3 Ultra · 512 GBQ4_K_LRuns great24 GB25
M2 · 8 GBIQ2_XSRuns well5.2 GB5.3
M2 · 16 GBQ4_K_LRuns well9.8 GB3.8
M2 · 24 GBQ4_K_LRuns great11 GB3.8
M2 Pro · 16 GBQ4_K_LRuns well9.8 GB7.3
M2 Pro · 32 GBQ4_K_LRuns great13 GB7.3
M2 Max · 32 GBQ4_K_LRuns great13 GB14
M2 Max · 64 GBQ4_K_LRuns great24 GB14
M2 Max · 96 GBQ4_K_LRuns great24 GB14
M2 Ultra · 64 GBQ4_K_LRuns great24 GB25
M2 Ultra · 128 GBQ4_K_LRuns great24 GB25
M2 Ultra · 192 GBQ4_K_LRuns great24 GB25
M1 · 8 GBIQ2_XSRuns well5.2 GB3.6
M1 · 16 GBQ4_K_LRuns well9.8 GB2.6
M1 Pro · 16 GBQ4_K_LRuns well9.8 GB7
M1 Pro · 32 GBQ4_K_LRuns great13 GB7
M1 Max · 32 GBQ4_K_LRuns great13 GB13
M1 Max · 64 GBQ4_K_LRuns great24 GB13
M1 Ultra · 64 GBQ4_K_LRuns great24 GB22
M1 Ultra · 128 GBQ4_K_LRuns great24 GB22

Writes is words a second, estimated from the Mac's memory speed and the model's size; how, and how wrong.

Packaged by

3 repositories on Hugging Face carry this model for llama.cpp. Outmake reads them all as one model.

Information

Maker
Microsoft
Released
May 2024
Parameters
14B
Longest chat
128K tokens
Reads pictures
No
Licence
mit
Downloads this month
12,079
Likes
66
Sizes
19, 3.0 to 8.6 GB
Kind
Maker's release

Badge for the model card

Shows on Phi 3 Medium 128K's Hugging Face card which Macs run it well, linked to this page. It updates every morning.

Runs on Mac badge

Files, downloads and likes read from Hugging Face on November 5, 2026 and refreshed every morning. Memory and speed are estimates until someone running Outmake measures this model on that Mac.

Your Computer

Every model is judged for this Mac. Both are under About This Mac, in the Apple menu.

Every Mac · Every graphics card

Civitai

Open in OutmakeHugging Face

New to Outmake? Download the free beta