Outmake Models
Models

Qwen2.5 3B

Alibaba Qwen · September 2024

Get Hugging Face

A 3.1B chat model from Alibaba Qwen.

Downloads
22K
Likes
12
Parameters
3.1B
Context
32K
Licence
other

Runs great on M4 · 16 GB

The Q8_0 size from bartowski is the one to get: a 3.3 GB download that uses 4.1 GB of memory, and writes about 16 words a second.

Ratings

KnowledgeMMLU-Pro
43.7%

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.

Qwen2.5 3B 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 bartowski/Qwen2.5-3B-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
IQ3_M1.5 GB1.8 GBestimatedOpen
Q3_K_L1.7 GB2.0 GBestimatedOpen
IQ4_XS1.7 GB2.0 GBestimatedOpen
Q3_K_XL1.8 GB2.1 GBestimatedOpen
Q4_0_4_41.8 GB2.1 GBestimatedOpen
Q4_0_4_81.8 GB2.1 GBestimatedOpen
Q4_0_8_81.8 GB2.1 GBestimatedOpen
Q4_01.8 GB2.1 GBestimatedOpen
Q4_K_S1.8 GB2.1 GBestimatedOpen
Q4_K_M1.9 GB2.2 GBestimatedOpen
Q4_K_L2.0 GB2.3 GBestimatedOpen
Q5_K_S2.2 GB2.5 GBestimatedOpen
Q5_K_M2.2 GB2.5 GBestimatedOpen
Q5_K_L2.3 GB2.6 GBestimatedOpen
Q6_K2.5 GB2.8 GBestimatedOpen
Q6_K_L2.6 GB2.9 GBestimatedOpen
Q8_03.3 GB3.6 GBestimatedOpen
F3212 GB13 GBestimatedOpen

Run It Anywhere

Qwen2.5 3B is an open model, so any app that runs GGUF files can run it. These take the Q8_0 size from bartowski straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/bartowski/Qwen2.5-3B-GGUF:Q8_0
  • llama.cppMac, Windows, Linuxllama-server -hf bartowski/Qwen2.5-3B-GGUF:Q8_0
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/bartowski/Qwen2.5-3B-GGUF:Q8_0
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBQ8_03.3 GB3.6 GB
8 GBQ8_03.3 GB3.6 GB
10 GBQ8_03.3 GB3.6 GB
11 GBQ8_03.3 GB3.6 GB
12 GBQ8_03.3 GB3.6 GB
16 GBF3212 GB13 GB
20 GBF3212 GB13 GB
24 GBF3212 GB13 GB
32 GBF3212 GB13 GB
48 GBF3212 GB13 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 GBQ8_0Runs great4.1 GB24
M6 · 24 GBQ8_0Runs great4.1 GB24
M6 · 32 GBQ8_0Runs great4.1 GB24
M5 · 16 GBQ8_0Runs great4.1 GB22
M5 · 24 GBQ8_0Runs great4.1 GB22
M5 · 32 GBQ8_0Runs great4.1 GB22
M5 Pro · 24 GBQ8_0Runs great4.1 GB43
M5 Pro · 48 GBQ8_0Runs great4.1 GB43
M5 Pro · 64 GBQ8_0Runs great4.1 GB43
M5 Max · 36 GBQ8_0Runs great4.1 GB62
M5 Max · 48 GBQ8_0Runs great4.1 GB79
M5 Max · 64 GBQ8_0Runs great4.1 GB79
M5 Max · 128 GBQ8_0Runs great4.1 GB79
M5 Ultra · 96 GBQ8_0Runs great4.1 GB138
M5 Ultra · 256 GBQ8_0Runs great4.1 GB138
M5 Ultra · 512 GBQ8_0Runs great4.1 GB138
M4 · 16 GBQ8_0Runs great4.1 GB16
M4 · 24 GBQ8_0Runs great4.1 GB16
M4 · 32 GBQ8_0Runs great4.1 GB16
M4 Pro · 24 GBQ8_0Runs great4.1 GB32
M4 Pro · 48 GBQ8_0Runs great4.1 GB32
M4 Pro · 64 GBQ8_0Runs great4.1 GB32
M4 Max · 36 GBQ8_0Runs great4.1 GB42
M4 Max · 48 GBQ8_0Runs great4.1 GB51
M4 Max · 64 GBQ8_0Runs great4.1 GB51
M4 Max · 96 GBQ8_0Runs great4.1 GB42
M4 Max · 128 GBQ8_0Runs great4.1 GB51
M3 · 8 GBQ8_0Runs great3.7 GB14
M3 · 16 GBQ8_0Runs great4.1 GB14
M3 · 24 GBQ8_0Runs great4.1 GB14
M3 Pro · 18 GBQ8_0Runs great4.1 GB20
M3 Pro · 36 GBQ8_0Runs great4.1 GB20
M3 Max · 36 GBQ8_0Runs great4.1 GB35
M3 Max · 48 GBQ8_0Runs great4.1 GB44
M3 Max · 64 GBQ8_0Runs great4.1 GB44
M3 Max · 96 GBQ8_0Runs great4.1 GB35
M3 Max · 128 GBQ8_0Runs great4.1 GB44
M3 Ultra · 96 GBQ8_0Runs great4.1 GB69
M3 Ultra · 256 GBQ8_0Runs great4.1 GB69
M3 Ultra · 512 GBQ8_0Runs great4.1 GB69
M2 · 8 GBQ8_0Runs great3.7 GB14
M2 · 16 GBQ8_0Runs great4.1 GB14
M2 · 24 GBQ8_0Runs great4.1 GB14
M2 Pro · 16 GBQ8_0Runs great4.1 GB25
M2 Pro · 32 GBQ8_0Runs great4.1 GB25
M2 Max · 32 GBQ8_0Runs great4.1 GB44
M2 Max · 64 GBQ8_0Runs great4.1 GB44
M2 Max · 96 GBQ8_0Runs great4.1 GB44
M2 Ultra · 64 GBQ8_0Runs great4.1 GB68
M2 Ultra · 128 GBQ8_0Runs great4.1 GB68
M2 Ultra · 192 GBQ8_0Runs great4.1 GB68
M1 · 8 GBQ8_0Runs great3.7 GB9.3
M1 · 16 GBQ8_0Runs great4.1 GB9.3
M1 Pro · 16 GBQ8_0Runs great4.1 GB23
M1 Pro · 32 GBQ8_0Runs great4.1 GB23
M1 Max · 32 GBQ8_0Runs great4.1 GB37
M1 Max · 64 GBQ8_0Runs great4.1 GB37
M1 Ultra · 64 GBQ8_0Runs great4.1 GB53
M1 Ultra · 128 GBQ8_0Runs great4.1 GB53

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

Packaged by

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

Information

Maker
Alibaba Qwen
Released
September 2024
Parameters
3.1B
Longest chat
32K tokens
Reads pictures
No
Licence
other
Downloads this month
22,411
Likes
12
Sizes
18, 1.5 to 12 GB
Kind
Maker's release

Badge for the model card

Shows on Qwen2.5 3B'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