Outmake Models
Models

Translategemma 12B

Google · January 2026

Get Hugging Face

A 12B chat model from Google that reads pictures as well as text.

Downloads
75K
Likes
35
Parameters
12B
Context
128K
Licence
gemma

Runs great on M4 · 16 GB

The Q4_K_S size from mradermacher is the one to get: a 6.9 GB download that uses 7.5 GB of memory, and writes about 7.2 words a second.

Translategemma 12B 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 mradermacher/translategemma-12b-it-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
Q2_K4.8 GB5.4 GBestimatedOpen
Q3_K_S5.5 GB6.0 GBestimatedOpen
Q3_K_M6.0 GB6.6 GBestimatedOpen
Q3_K_L6.5 GB7.1 GBestimatedOpen
IQ4_XS6.6 GB7.2 GBestimatedOpen
Q4_K_S6.9 GB7.5 GBestimatedOpen
Q4_K_M7.3 GB7.9 GBestimatedOpen
Q5_K_S8.2 GB8.8 GBestimatedOpen
Q5_K_M8.4 GB9.0 GBestimatedOpen
Q6_K9.7 GB10 GBestimatedOpen
Q8_013 GB13 GBestimatedOpen

Run It Anywhere

Translategemma 12B is an open model, so any app that runs GGUF files can run it. These take the Q4_K_S size from mradermacher straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/mradermacher/translategemma-12b-it-GGUF:Q4_K_S
  • llama.cppMac, Windows, Linuxllama-server -hf mradermacher/translategemma-12b-it-GGUF:Q4_K_S
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/mradermacher/translategemma-12b-it-GGUF:Q4_K_S
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBNone worth runningneeds 6.0 GB
8 GBQ4_K_M7.3 GB7.9 GB
10 GBQ5_K_M8.4 GB9.0 GB
11 GBQ6_K9.7 GB10 GB
12 GBQ6_K9.7 GB10 GB
16 GBQ8_013 GB13 GB
20 GBQ8_013 GB13 GB
24 GBQ8_013 GB13 GB
32 GBQ8_013 GB13 GB
48 GBQ8_013 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 GBQ4_K_SRuns great7.5 GB10
M6 · 24 GBQ6_KRuns great11 GB8.4
M6 · 32 GBQ8_0Runs great15 GB7.7
M5 · 16 GBQ4_K_SRuns great7.5 GB9.4
M5 · 24 GBQ6_KRuns great11 GB7.6
M5 · 32 GBQ8_0Runs great15 GB7
M5 Pro · 24 GBQ6_KRuns great11 GB15
M5 Pro · 48 GBQ8_0Runs great18 GB14
M5 Pro · 64 GBQ8_0Runs great18 GB14
M5 Max · 36 GBQ8_0Runs great15 GB20
M5 Max · 48 GBQ8_0Runs great18 GB27
M5 Max · 64 GBQ8_0Runs great18 GB27
M5 Max · 128 GBQ8_0Runs great18 GB27
M5 Ultra · 96 GBQ8_0Runs great18 GB50
M5 Ultra · 256 GBQ8_0Runs great18 GB50
M5 Ultra · 512 GBQ8_0Runs great18 GB50
M4 · 16 GBQ4_K_SRuns great7.5 GB7.2
M4 · 24 GBQ6_KRuns great11 GB5.7
M4 · 32 GBQ8_0Runs great15 GB5.3
M4 Pro · 24 GBQ6_KRuns great11 GB12
M4 Pro · 48 GBQ8_0Runs great18 GB11
M4 Pro · 64 GBQ8_0Runs great18 GB11
M4 Max · 36 GBQ8_0Runs great15 GB16
M4 Max · 48 GBQ8_0Runs great18 GB20
M4 Max · 64 GBQ8_0Runs great18 GB20
M4 Max · 96 GBQ8_0Runs great18 GB16
M4 Max · 128 GBQ8_0Runs great18 GB20
M3 · 8 GBQ2_KTight fit5.7 GB8
M3 · 16 GBQ4_K_SRuns great7.5 GB6.2
M3 · 24 GBQ6_KRuns great11 GB4.8
M3 Pro · 18 GBQ4_K_MRuns great8.2 GB8.7
M3 Pro · 36 GBQ8_0Runs great15 GB6.6
M3 Max · 36 GBQ8_0Runs great15 GB12
M3 Max · 48 GBQ8_0Runs great18 GB16
M3 Max · 64 GBQ8_0Runs great18 GB16
M3 Max · 96 GBQ8_0Runs great18 GB12
M3 Max · 128 GBQ8_0Runs great18 GB16
M3 Ultra · 96 GBQ8_0Runs great18 GB28
M3 Ultra · 256 GBQ8_0Runs great18 GB28
M3 Ultra · 512 GBQ8_0Runs great18 GB28
M2 · 8 GBQ2_KTight fit5.7 GB8
M2 · 16 GBQ4_K_SRuns great7.5 GB6.2
M2 · 24 GBQ6_KRuns great11 GB4.8
M2 Pro · 16 GBQ4_K_SRuns great7.5 GB12
M2 Pro · 32 GBQ8_0Runs great15 GB8.6
M2 Max · 32 GBQ8_0Runs great15 GB16
M2 Max · 64 GBQ8_0Runs great18 GB16
M2 Max · 96 GBQ8_0Runs great18 GB16
M2 Ultra · 64 GBQ8_0Runs great18 GB28
M2 Ultra · 128 GBQ8_0Runs great18 GB28
M2 Ultra · 192 GBQ8_0Runs great18 GB28
M1 · 8 GBQ2_KTight fit5.7 GB5.4
M1 · 16 GBQ4_K_SRuns great7.5 GB4.2
M1 Pro · 16 GBQ4_K_SRuns great7.5 GB11
M1 Pro · 32 GBQ8_0Runs great15 GB8.2
M1 Max · 32 GBQ8_0Runs great15 GB15
M1 Max · 64 GBQ8_0Runs great18 GB15
M1 Ultra · 64 GBQ8_0Runs great18 GB24
M1 Ultra · 128 GBQ8_0Runs great18 GB24

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
Google
Released
January 2026
Parameters
12B
Longest chat
128K tokens
Reads pictures
Yes
Licence
gemma
Downloads this month
74,894
Likes
35
Sizes
11, 4.8 to 13 GB
Kind
Maker's release

Badge for the model card

Shows on Translategemma 12B'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