Outmake Models
Models

Granite 4.1 8B

IBM · April 2026

Get Hugging Face

A 8.8B chat model from IBM.

Rating
2.9
Downloads
18K
Likes
65
Parameters
8.8B
Context
128K
Licence
apache-2.0

Runs great on M4 · 16 GB

The Q5_1 size from ibm-granite is the one to get: a 6.6 GB download that uses 7.5 GB of memory, and writes about 6.4 words a second.

Ratings

2.9 out of 5, across people's blind votes on its answers and the tests below.

Text
1304
Coding
1353

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.

KnowledgeMMLU-Pro
56%
ScienceGPQA Diamond
42%

MMLU-Pro: graduate-level questions in 14 subjects. GPQA Diamond: questions PhD scientists write for each other. Scores are the ones filed on the model's Hugging Face page or the board's own table, read 2026-11-05; some come from the maker.

Granite 4.1 8B 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 ibm-granite/granite-4.1-8b-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
Q2_K3.4 GB4.3 GBestimatedOpen
Q3_K_S3.9 GB4.8 GBestimatedOpen
Q3_K_M4.3 GB5.2 GBestimatedOpen
Q3_K_L4.7 GB5.6 GBestimatedOpen
Q4_05.1 GB5.9 GBestimatedOpen
Q4_K_S5.1 GB6.0 GBestimatedOpen
Q4_K_M5.3 GB6.2 GBestimatedOpen
Q4_15.6 GB6.5 GBestimatedOpen
Q5_06.1 GB7.0 GBestimatedOpen
Q5_K_S6.1 GB7.0 GBestimatedOpen
Q5_K_M6.3 GB7.1 GBestimatedOpen
Q5_16.6 GB7.5 GBestimatedOpen
Q6_K7.2 GB8.1 GBestimatedOpen
Q8_09.3 GB10 GBestimatedOpen
BF1618 GB18 GBestimatedOpen

Run It Anywhere

Granite 4.1 8B is an open model, so any app that runs GGUF files can run it. These take the Q5_1 size from ibm-granite straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/ibm-granite/granite-4.1-8b-GGUF:Q5_1
  • llama.cppMac, Windows, Linuxllama-server -hf ibm-granite/granite-4.1-8b-GGUF:Q5_1
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/ibm-granite/granite-4.1-8b-GGUF:Q5_1
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBQ3_K_L4.7 GB5.6 GB
8 GBQ5_16.6 GB7.5 GB
10 GBQ6_K7.2 GB8.1 GB
11 GBQ8_09.3 GB10 GB
12 GBQ8_09.3 GB10 GB
16 GBQ8_09.3 GB10 GB
20 GBBF1618 GB18 GB
24 GBBF1618 GB18 GB
32 GBBF1618 GB18 GB
48 GBBF1618 GB18 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_1Runs great7.5 GB9.2
M6 · 24 GBQ8_0Runs great11 GB7.7
M6 · 32 GBQ8_0Runs great13 GB7.7
M5 · 16 GBQ5_1Runs great7.5 GB8.3
M5 · 24 GBQ8_0Runs great11 GB7
M5 · 32 GBQ8_0Runs great13 GB7
M5 Pro · 24 GBQ8_0Runs great11 GB14
M5 Pro · 48 GBQ8_0Runs great22 GB14
M5 Pro · 64 GBQ8_0Runs great22 GB14
M5 Max · 36 GBQ8_0Runs great16 GB21
M5 Max · 48 GBQ8_0Runs great22 GB27
M5 Max · 64 GBQ8_0Runs great22 GB27
M5 Max · 128 GBQ8_0Runs great22 GB27
M5 Ultra · 96 GBQ8_0Runs great22 GB51
M5 Ultra · 256 GBQ8_0Runs great22 GB51
M5 Ultra · 512 GBQ8_0Runs great22 GB51
M4 · 16 GBQ5_1Runs great7.5 GB6.4
M4 · 24 GBQ8_0Runs great11 GB5.3
M4 · 32 GBQ8_0Runs great13 GB5.3
M4 Pro · 24 GBQ8_0Runs great11 GB11
M4 Pro · 48 GBQ8_0Runs great22 GB11
M4 Pro · 64 GBQ8_0Runs great22 GB11
M4 Max · 36 GBQ8_0Runs great16 GB16
M4 Max · 48 GBQ8_0Runs great22 GB21
M4 Max · 64 GBQ8_0Runs great22 GB21
M4 Max · 96 GBQ8_0Runs great22 GB16
M4 Max · 128 GBQ8_0Runs great22 GB21
M3 · 8 GBQ3_K_MRuns well5.2 GB5.9
M3 · 16 GBQ5_1Runs great7.5 GB5.5
M3 · 24 GBQ8_0Runs great11 GB4.5
M3 Pro · 18 GBQ6_KRuns great8.1 GB6.8
M3 Pro · 36 GBQ8_0Runs great16 GB6.6
M3 Max · 36 GBQ8_0Runs great16 GB13
M3 Max · 48 GBQ8_0Runs great22 GB16
M3 Max · 64 GBQ8_0Runs great22 GB16
M3 Max · 96 GBQ8_0Runs great22 GB13
M3 Max · 128 GBQ8_0Runs great22 GB16
M3 Ultra · 96 GBQ8_0Runs great22 GB29
M3 Ultra · 256 GBQ8_0Runs great22 GB29
M3 Ultra · 512 GBQ8_0Runs great22 GB29
M2 · 8 GBQ3_K_MRuns well5.2 GB5.9
M2 · 16 GBQ5_1Runs great7.5 GB5.5
M2 · 24 GBQ8_0Runs great11 GB4.5
M2 Pro · 16 GBQ5_1Runs great7.5 GB11
M2 Pro · 32 GBQ8_0Runs great13 GB8.7
M2 Max · 32 GBQ8_0Runs great13 GB16
M2 Max · 64 GBQ8_0Runs great22 GB16
M2 Max · 96 GBQ8_0Runs great22 GB16
M2 Ultra · 64 GBQ8_0Runs great22 GB29
M2 Ultra · 128 GBQ8_0Runs great22 GB29
M2 Ultra · 192 GBQ8_0Runs great22 GB29
M1 · 8 GBQ3_K_MRuns well5.2 GB4
M1 · 16 GBQ5_1Runs great7.5 GB3.7
M1 Pro · 16 GBQ5_1Runs great7.5 GB10
M1 Pro · 32 GBQ8_0Runs great13 GB8.4
M1 Max · 32 GBQ8_0Runs great13 GB15
M1 Max · 64 GBQ8_0Runs great22 GB15
M1 Ultra · 64 GBQ8_0Runs great22 GB26
M1 Ultra · 128 GBQ8_0Runs great22 GB26

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

Packaged by

2 repositories on Hugging Face carry this model for llama.cpp. Outmake reads them all as one model.

Information

Maker
IBM
Released
April 2026
Parameters
8.8B
Longest chat
128K tokens
Reads pictures
No
Licence
apache-2.0
Downloads this month
17,701
Likes
65
Sizes
15, 3.4 to 18 GB
Kind
Maker's release

Badge for the model card

Shows on Granite 4.1 8B'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