Outmake Models
Models

MiniCPM5 2B

openbmb · September 2026

Get Hugging Face

A 2.5B chat model from openbmb.

Downloads
373K
Likes
426
Parameters
2.5B
Context
128K
Licence
apache-2.0

Runs great on M4 · 16 GB

The Q5_K_L size from Mungert is the one to get: a 2.0 GB download that uses 5.4 GB of memory, and writes about 17 words a second.

MiniCPM5 2B 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/MiniCPM5-2B-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
IQ2_M1.0 GB1.3 GBestimatedOpen
Q2_K1.0 GB1.3 GBestimatedOpen
IQ3_XXS1.1 GB1.5 GBestimatedOpen
Q3_K_S1.2 GB1.5 GBestimatedOpen
IQ3_XS1.2 GB1.5 GBestimatedOpen
Q3_K_M1.2 GB1.5 GBestimatedOpen
Q3_K_L1.3 GB1.6 GBestimatedOpen
IQ3_M1.4 GB1.7 GBestimatedOpen
IQ4_XS1.5 GB1.8 GBestimatedOpen
Q4_01.5 GB1.8 GBestimatedOpen
Q4_K_S1.5 GB1.8 GBestimatedOpen
IQ4_NL1.6 GB1.9 GBestimatedOpen
Q4_11.6 GB2.0 GBestimatedOpen
Q4_K_L1.7 GB2.0 GBestimatedOpen
Q5_K_S1.8 GB2.1 GBestimatedOpen
Q5_K_M1.9 GB2.2 GBestimatedOpen

Run It Anywhere

MiniCPM5 2B is an open model, so any app that runs GGUF files can run it. These take the Q5_K_L size from Mungert straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/Mungert/MiniCPM5-2B-GGUF:Q5_K_L
  • llama.cppMac, Windows, Linuxllama-server -hf Mungert/MiniCPM5-2B-GGUF:Q5_K_L
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/Mungert/MiniCPM5-2B-GGUF:Q5_K_L
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBQ5_K_M1.9 GB2.2 GB
8 GBQ5_K_M1.9 GB2.2 GB
10 GBQ5_K_M1.9 GB2.2 GB
11 GBQ5_K_M1.9 GB2.2 GB
12 GBQ5_K_M1.9 GB2.2 GB
16 GBQ5_K_M1.9 GB2.2 GB
20 GBQ5_K_M1.9 GB2.2 GB
24 GBQ5_K_M1.9 GB2.2 GB
32 GBQ5_K_M1.9 GB2.2 GB
48 GBQ5_K_M1.9 GB2.2 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 GBQ5_K_LRuns great5.4 GB26
M6 · 24 GBQ5_K_LRuns great5.4 GB26
M6 · 32 GBQ5_K_LRuns great5.4 GB26
M5 · 16 GBQ5_K_LRuns great5.4 GB24
M5 · 24 GBQ5_K_LRuns great5.4 GB24
M5 · 32 GBQ5_K_LRuns great5.4 GB24
M5 Pro · 24 GBQ5_K_LRuns great5.4 GB46
M5 Pro · 48 GBQ5_K_LRuns great5.4 GB46
M5 Pro · 64 GBQ5_K_LRuns great5.4 GB46
M5 Max · 36 GBQ5_K_LRuns great5.4 GB65
M5 Max · 48 GBQ5_K_LRuns great5.4 GB83
M5 Max · 64 GBQ5_K_LRuns great5.4 GB83
M5 Max · 128 GBQ5_K_LRuns great5.4 GB83
M5 Ultra · 96 GBQ5_K_LRuns great5.4 GB136
M5 Ultra · 256 GBQ5_K_LRuns great5.4 GB136
M5 Ultra · 512 GBQ5_K_LRuns great5.4 GB136
M4 · 16 GBQ5_K_LRuns great5.4 GB17
M4 · 24 GBQ5_K_LRuns great5.4 GB17
M4 · 32 GBQ5_K_LRuns great5.4 GB17
M4 Pro · 24 GBQ5_K_LRuns great5.4 GB32
M4 Pro · 48 GBQ5_K_LRuns great5.4 GB32
M4 Pro · 64 GBQ5_K_LRuns great5.4 GB32
M4 Max · 36 GBQ5_K_LRuns great5.4 GB42
M4 Max · 48 GBQ5_K_LRuns great5.4 GB49
M4 Max · 64 GBQ5_K_LRuns great5.4 GB49
M4 Max · 96 GBQ5_K_LRuns great5.4 GB42
M4 Max · 128 GBQ5_K_LRuns great5.4 GB49
M3 · 8 GBQ5_K_LRuns great3.8 GB15
M3 · 16 GBQ5_K_LRuns great5.4 GB15
M3 · 24 GBQ5_K_LRuns great5.4 GB15
M3 Pro · 18 GBQ5_K_LRuns great5.4 GB21
M3 Pro · 36 GBQ5_K_LRuns great5.4 GB21
M3 Max · 36 GBQ5_K_LRuns great5.4 GB36
M3 Max · 48 GBQ5_K_LRuns great5.4 GB44
M3 Max · 64 GBQ5_K_LRuns great5.4 GB44
M3 Max · 96 GBQ5_K_LRuns great5.4 GB36
M3 Max · 128 GBQ5_K_LRuns great5.4 GB44
M3 Ultra · 96 GBQ5_K_LRuns great5.4 GB64
M3 Ultra · 256 GBQ5_K_LRuns great5.4 GB64
M3 Ultra · 512 GBQ5_K_LRuns great5.4 GB64
M2 · 8 GBQ5_K_LRuns great3.8 GB15
M2 · 16 GBQ5_K_LRuns great5.4 GB15
M2 · 24 GBQ5_K_LRuns great5.4 GB15
M2 Pro · 16 GBQ5_K_LRuns great5.4 GB27
M2 Pro · 32 GBQ5_K_LRuns great5.4 GB27
M2 Max · 32 GBQ5_K_LRuns great5.4 GB44
M2 Max · 64 GBQ5_K_LRuns great5.4 GB44
M2 Max · 96 GBQ5_K_LRuns great5.4 GB44
M2 Ultra · 64 GBQ5_K_LRuns great5.4 GB64
M2 Ultra · 128 GBQ5_K_LRuns great5.4 GB64
M2 Ultra · 192 GBQ5_K_LRuns great5.4 GB64
M1 · 8 GBQ5_K_LRuns great3.8 GB10
M1 · 16 GBQ5_K_LRuns great5.4 GB10
M1 Pro · 16 GBQ5_K_LRuns great5.4 GB23
M1 Pro · 32 GBQ5_K_LRuns great5.4 GB23
M1 Max · 32 GBQ5_K_LRuns great5.4 GB35
M1 Max · 64 GBQ5_K_LRuns great5.4 GB35
M1 Ultra · 64 GBQ5_K_LRuns great5.4 GB47
M1 Ultra · 128 GBQ5_K_LRuns great5.4 GB47

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
openbmb
Released
September 2026
Parameters
2.5B
Longest chat
128K tokens
Reads pictures
No
Licence
apache-2.0
Downloads this month
372,742
Likes
426
Sizes
16, 1.0 to 1.9 GB
Kind
Maker's release

Badge for the model card

Shows on MiniCPM5 2B'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