Outmake Models
Models

Gemma 2 27B

Google · June 2024

Get Hugging Face

A 27B chat model from Google.

Rating
2.9
Downloads
23K
Likes
211
Parameters
27B
Context
8K
Licence
gemma

Runs great on M4 · 16 GB

The IQ1_S size from legraphista is the one to get: a 6.1 GB download that uses 7.5 GB of memory, and writes about 5.5 words a second.

Ratings

2.9 out of 5, across people's blind votes on its answers.

Text
1289
Coding
1304

LM Arena ratings of October 28, 2026 (CC BY 4.0): people vote between two answers without knowing which model wrote which. The bar ends at 1488, the best any open model has. The Arena tests the full model; a small size on a Mac answers a little worse.

Gemma 2 27B 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 bartowski/gemma-2-27b-it-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
IQ2_XS8.4 GB9.8 GBestimatedOpen
IQ2_S8.7 GB10 GBestimatedOpen
IQ2_M9.4 GB11 GBestimatedOpen
Q2_K10 GB12 GBestimatedOpen
Q2_K_L11 GB12 GBestimatedOpen
IQ3_XXS11 GB12 GBestimatedOpen
IQ3_XS12 GB13 GBestimatedOpen
Q3_K_S12 GB14 GBestimatedOpen
IQ3_M12 GB14 GBestimatedOpen
Q3_K_M13 GB15 GBestimatedOpen
Q3_K_L15 GB16 GBestimatedOpen
Q3_K_XL15 GB16 GBestimatedOpen
IQ4_XS15 GB16 GBestimatedOpen
Q4_K_S16 GB17 GBestimatedOpen
Q4_K_M17 GB18 GBestimatedOpen
Q4_K_L17 GB18 GBestimatedOpen
Q5_K_S19 GB20 GBestimatedOpen
Q5_K_M19 GB21 GBestimatedOpen
Q5_K_L20 GB21 GBestimatedOpen
Q6_K22 GB24 GBestimatedOpen
Q6_K_L23 GB24 GBestimatedOpen
Q8_029 GB30 GBestimatedOpen
Q8_0_L30 GB31 GBestimatedOpen

Run It Anywhere

Gemma 2 27B is an open model, so any app that runs GGUF files can run it. These take the IQ1_S size from legraphista straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/legraphista/gemma-2-27b-it-IMat-GGUF:IQ1_S
  • llama.cppMac, Windows, Linuxllama-server -hf legraphista/gemma-2-27b-it-IMat-GGUF:IQ1_S
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/legraphista/gemma-2-27b-it-IMat-GGUF:IQ1_S
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBNone worth runningneeds 12 GB
8 GBNone worth runningneeds 12 GB
10 GBNone worth runningneeds 12 GB
11 GBNone worth runningneeds 12 GB
12 GBNone worth runningneeds 12 GB
16 GBQ3_K_M13 GB15 GB
20 GBQ4_K_L17 GB18 GB
24 GBQ5_K_L20 GB21 GB
32 GBQ8_0_L30 GB31 GB
48 GBQ8_0_L30 GB31 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 GBIQ1_SRuns great7.5 GB7.9
M6 · 24 GBQ3_K_LRuns well16 GB4.7
M6 · 32 GBQ3_K_MRuns great15 GB4.9
M5 · 16 GBIQ1_SRuns great7.5 GB7.1
M5 · 24 GBQ3_K_LRuns well16 GB4.2
M5 · 32 GBQ3_K_MRuns great15 GB4.5
M5 Pro · 24 GBQ3_K_LRuns well16 GB8.4
M5 Pro · 48 GBQ5_K_LRuns great21 GB6.7
M5 Pro · 64 GBQ8_0Runs great30 GB5.9
M5 Max · 36 GBQ4_K_SRuns great17 GB12
M5 Max · 48 GBQ5_K_LRuns great21 GB13
M5 Max · 64 GBQ8_0Runs great30 GB12
M5 Max · 128 GBQ8_0Runs great30 GB12
M5 Ultra · 96 GBQ8_0Runs great30 GB23
M5 Ultra · 256 GBQ8_0Runs great30 GB23
M5 Ultra · 512 GBQ8_0Runs great30 GB23
M4 · 16 GBIQ1_SRuns great7.5 GB5.5
M4 · 24 GBQ3_K_LRuns well16 GB3.3
M4 · 32 GBQ3_K_MRuns great15 GB3.5
M4 Pro · 24 GBQ3_K_LRuns well16 GB7.3
M4 Pro · 48 GBQ5_K_LRuns great21 GB5.9
M4 Pro · 64 GBQ8_0Runs great30 GB5
M4 Max · 36 GBQ4_K_SRuns great17 GB10
M4 Max · 48 GBQ5_K_LRuns great21 GB11
M4 Max · 64 GBQ8_0Runs great30 GB9.6
M4 Max · 96 GBQ8_0Runs great30 GB7.4
M4 Max · 128 GBQ8_0Runs great30 GB9.6
M3 · 8 GBIQ1_SToo big7.5 GB
M3 · 16 GBIQ1_SRuns great7.5 GB4.7
M3 · 24 GBQ3_K_LRuns well16 GB2.8
M3 Pro · 18 GBQ2_KRuns well12 GB5.1
M3 Pro · 36 GBQ4_K_SRuns great17 GB3.9
M3 Max · 36 GBQ4_K_SRuns great17 GB7.6
M3 Max · 48 GBQ5_K_LRuns great21 GB8.5
M3 Max · 64 GBQ8_0Runs great30 GB7.3
M3 Max · 96 GBQ8_0Runs great30 GB5.5
M3 Max · 128 GBQ8_0Runs great30 GB7.3
M3 Ultra · 96 GBQ8_0Runs great30 GB14
M3 Ultra · 256 GBQ8_0Runs great30 GB14
M3 Ultra · 512 GBQ8_0Runs great30 GB14
M2 · 8 GBIQ1_SToo big7.5 GB
M2 · 16 GBIQ1_SRuns great7.5 GB4.7
M2 · 24 GBQ3_K_LRuns well16 GB2.8
M2 Pro · 16 GBIQ1_SRuns great7.5 GB8.9
M2 Pro · 32 GBQ3_K_MRuns great15 GB5.8
M2 Max · 32 GBQ3_K_MRuns great15 GB11
M2 Max · 64 GBQ8_0Runs great30 GB7.3
M2 Max · 96 GBQ8_0Runs great30 GB7.3
M2 Ultra · 64 GBQ8_0Runs great30 GB14
M2 Ultra · 128 GBQ8_0Runs great30 GB14
M2 Ultra · 192 GBQ8_0Runs great30 GB14
M1 · 8 GBIQ1_SToo big7.5 GB
M1 · 16 GBIQ1_SRuns great7.5 GB3.2
M1 Pro · 16 GBIQ1_SRuns great7.5 GB8.5
M1 Pro · 32 GBQ3_K_MRuns great15 GB5.6
M1 Max · 32 GBQ3_K_MRuns great15 GB10
M1 Max · 64 GBQ8_0Runs great30 GB7
M1 Ultra · 64 GBQ8_0Runs great30 GB13
M1 Ultra · 128 GBQ8_0Runs great30 GB13

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
June 2024
Parameters
27B
Longest chat
8K tokens
Reads pictures
No
Licence
gemma
Downloads this month
22,998
Likes
211
Sizes
23, 8.4 to 30 GB
Kind
Maker's release

Badge for the model card

Shows on Gemma 2 27B'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