Outmake Models
Models

Ternary Bonsai 27B

PrismML · July 2026

Get Hugging Face

A 27B chat model from PrismML that reads pictures as well as text.

Downloads
351K
Likes
1.4K
Parameters
27B
Context
256K
Licence
apache-2.0

Runs great on M4 · 16 GB

The PQ2_0 size from prism-ml is the one to get: a 7.2 GB download that uses 7.6 GB of memory, and writes about 7.7 words a second.

Ternary Bonsai 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 prism-ml/Ternary-Bonsai-27B-gguf on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
PQ2_07.2 GB7.6 GBestimatedOpen
Q2_07.2 GB7.6 GBestimatedOpen
Q2_G647.6 GB8.0 GBestimatedOpen
F1654 GB54 GBestimatedOpen

Run It Anywhere

Ternary Bonsai 27B is an open model, so any app that runs GGUF files can run it. These take the PQ2_0 size from prism-ml straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:PQ2_0
  • llama.cppMac, Windows, Linuxllama-server -hf prism-ml/Ternary-Bonsai-27B-gguf:PQ2_0
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:PQ2_0
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBNone worth runningneeds 54 GB
8 GBNone worth runningneeds 54 GB
10 GBNone worth runningneeds 54 GB
11 GBNone worth runningneeds 54 GB
12 GBNone worth runningneeds 54 GB
16 GBNone worth runningneeds 54 GB
20 GBNone worth runningneeds 54 GB
24 GBNone worth runningneeds 54 GB
32 GBNone worth runningneeds 54 GB
48 GBNone worth runningneeds 54 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 GBPQ2_0Runs great7.6 GB11
M6 · 24 GBQ2_G64Runs great10 GB11
M6 · 32 GBQ2_G64Runs great13 GB11
M5 · 16 GBPQ2_0Runs great7.6 GB10
M5 · 24 GBQ2_G64Runs great10 GB9.8
M5 · 32 GBQ2_G64Runs great13 GB9.8
M5 Pro · 24 GBQ2_G64Runs great10 GB19
M5 Pro · 48 GBQ2_G64Runs great18 GB19
M5 Pro · 64 GBQ2_G64Runs great18 GB19
M5 Max · 36 GBQ2_G64Runs great13 GB28
M5 Max · 48 GBQ2_G64Runs great18 GB36
M5 Max · 64 GBQ2_G64Runs great18 GB36
M5 Max · 128 GBF16Runs great64 GB7.6
M5 Ultra · 96 GBF16Runs well64 GB15
M5 Ultra · 256 GBF16Runs great64 GB15
M5 Ultra · 512 GBF16Runs great64 GB15
M4 · 16 GBPQ2_0Runs great7.6 GB7.7
M4 · 24 GBQ2_G64Runs great10 GB7.4
M4 · 32 GBQ2_G64Runs great13 GB7.4
M4 Pro · 24 GBQ2_G64Runs great10 GB15
M4 Pro · 48 GBQ2_G64Runs great18 GB15
M4 Pro · 64 GBQ2_G64Runs great18 GB15
M4 Max · 36 GBQ2_G64Runs great13 GB20
M4 Max · 48 GBQ2_G64Runs great18 GB25
M4 Max · 64 GBQ2_G64Runs great18 GB25
M4 Max · 96 GBF16Runs well64 GB4.7
M4 Max · 128 GBF16Runs great64 GB6.1
M3 · 8 GBPQ2_0Too big18 GB
M3 · 16 GBPQ2_0Runs great7.6 GB6.6
M3 · 24 GBQ2_G64Runs great10 GB6.4
M3 Pro · 18 GBQ2_G64Runs great8.3 GB9.2
M3 Pro · 36 GBQ2_G64Runs great13 GB9.2
M3 Max · 36 GBQ2_G64Runs great13 GB17
M3 Max · 48 GBQ2_G64Runs great18 GB21
M3 Max · 64 GBQ2_G64Runs great18 GB21
M3 Max · 96 GBF16Runs well64 GB3.5
M3 Max · 128 GBF16Runs great64 GB4.6
M3 Ultra · 96 GBF16Runs well64 GB8.8
M3 Ultra · 256 GBF16Runs great64 GB8.8
M3 Ultra · 512 GBF16Runs great64 GB8.8
M2 · 8 GBPQ2_0Too big18 GB
M2 · 16 GBPQ2_0Runs great7.6 GB6.6
M2 · 24 GBQ2_G64Runs great10 GB6.4
M2 Pro · 16 GBPQ2_0Runs great7.6 GB12
M2 Pro · 32 GBQ2_G64Runs great13 GB12
M2 Max · 32 GBQ2_G64Runs great13 GB21
M2 Max · 64 GBQ2_G64Runs great18 GB21
M2 Max · 96 GBF16Runs well64 GB4.6
M2 Ultra · 64 GBQ2_G64Runs great18 GB34
M2 Ultra · 128 GBF16Runs great64 GB8.8
M2 Ultra · 192 GBF16Runs great64 GB8.8
M1 · 8 GBPQ2_0Too big18 GB
M1 · 16 GBPQ2_0Runs great7.6 GB4.5
M1 Pro · 16 GBPQ2_0Runs great7.6 GB11
M1 Pro · 32 GBQ2_G64Runs great13 GB11
M1 Max · 32 GBQ2_G64Runs great13 GB18
M1 Max · 64 GBQ2_G64Runs great18 GB18
M1 Ultra · 64 GBQ2_G64Runs great18 GB27
M1 Ultra · 128 GBF16Runs great64 GB8.3

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

Information

Maker
PrismML
Released
July 2026
Parameters
27B
Longest chat
256K tokens
Reads pictures
Yes
Licence
apache-2.0
Downloads this month
351,142
Likes
1,437
Sizes
4, 7.2 to 54 GB
Kind
Maker's release

Badge for the model card

Shows on Ternary Bonsai 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