Outmake Models
Models

Mistral Small 3.2 24B 2506

Mistral AI · June 2025

Get Hugging Face

A 24B chat model from Mistral AI that reads pictures as well as text.

Downloads
193K
Likes
250
Parameters
24B
Context
128K
Licence
apache-2.0

Runs well on M4 · 16 GB

The UD-IQ3_XXS size from unsloth is the one to get: a 9.4 GB download that uses 10 GB of memory, and writes about 4.6 words a second.

Mistral Small 3.2 24B 2506 runs well from a Mac with an M1 and 16 GB of memory. 56 of 59 Mac configurations run it well.

Compare with

The nearest models in size.

Sizes

From unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
UD-IQ1_S5.6 GB6.4 GBestimatedOpen
UD-IQ1_M6.0 GB6.9 GBestimatedOpen
UD-IQ2_XXS6.8 GB7.6 GBestimatedOpen
UD-IQ2_M8.2 GB9.1 GBestimatedOpen
Q2_K8.9 GB9.8 GBestimatedOpen
Q2_K_L9.0 GB9.9 GBestimatedOpen
UD-Q2_K_XL9.3 GB10 GBestimatedOpen
UD-IQ3_XXS9.4 GB10 GBestimatedOpen
Q3_K_S10 GB11 GBestimatedOpen
Q3_K_M11 GB12 GBestimatedOpen
UD-Q3_K_XL12 GB13 GBestimatedOpen
IQ4_XS13 GB14 GBestimatedOpen
IQ4_NL13 GB14 GBestimatedOpen
Q4_013 GB14 GBestimatedOpen
Q4_K_S14 GB14 GBestimatedOpen
Q4_K_M14 GB15 GBestimatedOpen
UD-Q4_K_XL15 GB15 GBestimatedOpen
Q4_115 GB16 GBestimatedOpen
Q5_K_S16 GB17 GBestimatedOpen
Q5_K_M17 GB18 GBestimatedOpen
UD-Q5_K_XL17 GB18 GBestimatedOpen
Q6_K19 GB20 GBestimatedOpen
UD-Q6_K_XL21 GB22 GBestimatedOpen
Q8_025 GB26 GBestimatedOpen
UD-Q8_K_XL29 GB30 GBestimatedOpen
BF1647 GB48 GBestimatedOpen

Run It Anywhere

Mistral Small 3.2 24B 2506 is an open model, so any app that runs GGUF files can run it. These take the UD-IQ3_XXS size from unsloth straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF:UD-IQ3_XXS
  • llama.cppMac, Windows, Linuxllama-server -hf unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF:UD-IQ3_XXS
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/unsloth/Mistral-Small-3.2-24B-Instruct-2506-GGUF:UD-IQ3_XXS
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBNone worth runningneeds 9.1 GB
8 GBNone worth runningneeds 9.1 GB
10 GBUD-IQ2_M8.2 GB9.1 GB
11 GBUD-IQ3_XXS9.4 GB10 GB
12 GBQ3_K_S10 GB11 GB
16 GBQ4_115 GB16 GB
20 GBUD-Q5_K_XL17 GB18 GB
24 GBUD-Q6_K_XL21 GB22 GB
32 GBUD-Q8_K_XL29 GB30 GB
48 GBUD-Q8_K_XL29 GB30 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 GBUD-IQ3_XXSRuns well10 GB6.5
M6 · 24 GBQ4_1Runs well16 GB5.4
M6 · 32 GBUD-Q4_K_XLRuns great15 GB4.8
M5 · 16 GBUD-IQ3_XXSRuns well10 GB5.9
M5 · 24 GBQ4_1Runs well16 GB4.9
M5 · 32 GBUD-Q4_K_XLRuns great15 GB4.4
M5 Pro · 24 GBQ4_1Runs well16 GB9.7
M5 Pro · 48 GBUD-Q6_K_XLRuns great22 GB6.6
M5 Pro · 64 GBQ8_0Runs great28 GB6.8
M5 Max · 36 GBQ5_K_SRuns great17 GB12
M5 Max · 48 GBUD-Q6_K_XLRuns great22 GB13
M5 Max · 64 GBQ8_0Runs great28 GB13
M5 Max · 128 GBQ8_0Runs great38 GB13
M5 Ultra · 96 GBQ8_0Runs great38 GB26
M5 Ultra · 256 GBQ8_0Runs great38 GB26
M5 Ultra · 512 GBQ8_0Runs great38 GB26
M4 · 16 GBUD-IQ3_XXSRuns well10 GB4.6
M4 · 24 GBQ4_1Runs well16 GB3.9
M4 · 32 GBUD-Q4_K_XLRuns great15 GB3.5
M4 Pro · 24 GBQ4_1Runs well16 GB8.5
M4 Pro · 48 GBUD-Q6_K_XLRuns great22 GB5.9
M4 Pro · 64 GBQ8_0Runs great28 GB5.8
M4 Max · 36 GBQ5_K_SRuns great17 GB10
M4 Max · 48 GBUD-Q6_K_XLRuns great22 GB11
M4 Max · 64 GBQ8_0Runs great28 GB11
M4 Max · 96 GBQ8_0Runs great38 GB8.5
M4 Max · 128 GBQ8_0Runs great38 GB11
M3 · 8 GBUD-IQ1_SToo big18 GB
M3 · 16 GBUD-IQ3_XXSRuns well10 GB3.9
M3 · 24 GBQ4_1Runs well16 GB3.3
M3 Pro · 18 GBQ3_K_SRuns well11 GB5.4
M3 Pro · 36 GBQ5_K_SRuns great17 GB4
M3 Max · 36 GBQ5_K_SRuns great17 GB7.7
M3 Max · 48 GBUD-Q6_K_XLRuns great22 GB8.5
M3 Max · 64 GBQ8_0Runs great28 GB8.4
M3 Max · 96 GBQ8_0Runs great38 GB6.4
M3 Max · 128 GBQ8_0Runs great38 GB8.4
M3 Ultra · 96 GBQ8_0Runs great38 GB16
M3 Ultra · 256 GBQ8_0Runs great38 GB16
M3 Ultra · 512 GBQ8_0Runs great38 GB16
M2 · 8 GBUD-IQ1_SToo big18 GB
M2 · 16 GBUD-IQ3_XXSRuns well10 GB3.9
M2 · 24 GBQ4_1Runs well16 GB3.3
M2 Pro · 16 GBUD-IQ3_XXSRuns well10 GB7.6
M2 Pro · 32 GBUD-Q4_K_XLRuns great15 GB5.7
M2 Max · 32 GBUD-Q4_K_XLRuns great15 GB11
M2 Max · 64 GBQ8_0Runs great28 GB8.4
M2 Max · 96 GBQ8_0Runs great38 GB8.4
M2 Ultra · 64 GBQ8_0Runs great28 GB16
M2 Ultra · 128 GBQ8_0Runs great38 GB16
M2 Ultra · 192 GBQ8_0Runs great38 GB16
M1 · 8 GBUD-IQ1_SToo big18 GB
M1 · 16 GBUD-IQ3_XXSRuns well10 GB2.7
M1 Pro · 16 GBUD-IQ3_XXSRuns well10 GB7.3
M1 Pro · 32 GBUD-Q4_K_XLRuns great15 GB5.5
M1 Max · 32 GBUD-Q4_K_XLRuns great15 GB10
M1 Max · 64 GBQ8_0Runs great28 GB8.1
M1 Ultra · 64 GBQ8_0Runs great28 GB15
M1 Ultra · 128 GBQ8_0Runs great38 GB15

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
Mistral AI
Released
June 2025
Parameters
24B
Longest chat
128K tokens
Reads pictures
Yes
Licence
apache-2.0
Downloads this month
193,136
Likes
250
Sizes
26, 5.6 to 47 GB
Kind
Maker's release

Badge for the model card

Shows on Mistral Small 3.2 24B 2506'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