Outmake Models
Models

Llemma 7B

EleutherAI · February 2025

Get Hugging Face

A 6.7B chat model from EleutherAI.

Downloads
2.8K
Parameters
6.7B
Context
4K
Licence
llama2

Runs great on M4 · 16 GB

The Q6_K size from mradermacher is the one to get: a 5.5 GB download that uses 6.9 GB of memory, and writes about 3.1 words a second.

Ratings

KnowledgeMMLU-Pro
23.4%

MMLU-Pro: graduate-level questions in 14 subjects. 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.

Llemma 7B 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 mradermacher/llemma_7b-i1-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

SizeDownloadMemory at 8KGet
IQ1_S1.5 GB4.1 GBestimatedOpen
IQ1_M1.7 GB4.2 GBestimatedOpen
IQ2_XXS1.9 GB4.4 GBestimatedOpen
IQ2_XS2.0 GB4.6 GBestimatedOpen
IQ2_S2.2 GB4.8 GBestimatedOpen
Q2_K_S2.3 GB4.9 GBestimatedOpen
IQ2_M2.4 GB5.0 GBestimatedOpen
Q2_K2.5 GB5.1 GBestimatedOpen
IQ3_XXS2.6 GB5.2 GBestimatedOpen
IQ3_XS2.8 GB5.4 GBestimatedOpen
IQ3_S2.9 GB5.5 GBestimatedOpen
Q3_K_S2.9 GB5.5 GBestimatedOpen
IQ3_M3.1 GB5.7 GBestimatedOpen
Q3_K_M3.3 GB5.9 GBestimatedOpen
Q3_K_L3.6 GB6.2 GBestimatedOpen
IQ4_XS3.6 GB6.2 GBestimatedOpen
IQ4_NL3.8 GB6.4 GBestimatedOpen
Q4_03.8 GB6.4 GBestimatedOpen
Q4_K_S3.9 GB6.4 GBestimatedOpen
Q4_K_M4.1 GB6.7 GBestimatedOpen
Q4_14.2 GB6.8 GBestimatedOpen
Q5_K_S4.7 GB7.2 GBestimatedOpen
Q5_K_M4.8 GB7.4 GBestimatedOpen
Q6_K5.5 GB8.1 GBestimatedOpen

Run It Anywhere

Llemma 7B is an open model, so any app that runs GGUF files can run it. These take the Q6_K size from mradermacher straight from Hugging Face.

  • LM StudioMac, Windows, LinuxOpen
  • JanMac, Windows, LinuxOpen
  • MstyMac, Windows, LinuxOpen
  • OllamaMac, Windows, Linuxollama run hf.co/mradermacher/llemma_7b-i1-GGUF:Q6_K
  • llama.cppMac, Windows, Linuxllama-server -hf mradermacher/llemma_7b-i1-GGUF:Q6_K
  • Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/mradermacher/llemma_7b-i1-GGUF:Q6_K
On a PC Graphics Card
Video memoryLargest size that fitsDownloadMemory at 8K
6 GBQ3_K_M3.3 GB5.9 GB
8 GBQ5_K_M4.8 GB7.4 GB
10 GBQ6_K5.5 GB8.1 GB
11 GBQ6_K5.5 GB8.1 GB
12 GBQ6_K5.5 GB8.1 GB
16 GBQ6_K5.5 GB8.1 GB
20 GBQ6_K5.5 GB8.1 GB
24 GBQ6_K5.5 GB8.1 GB
32 GBQ6_K5.5 GB8.1 GB
48 GBQ6_K5.5 GB8.1 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 GBQ6_KRuns great6.9 GB4.5
M6 · 24 GBQ6_KRuns great6.9 GB4.5
M6 · 32 GBQ6_KRuns great6.9 GB4.5
M5 · 16 GBQ6_KRuns great6.9 GB4.1
M5 · 24 GBQ6_KRuns great6.9 GB4.1
M5 · 32 GBQ6_KRuns great6.9 GB4.1
M5 Pro · 24 GBQ6_KRuns great6.9 GB8.1
M5 Pro · 48 GBQ6_KRuns great6.9 GB8.1
M5 Pro · 64 GBQ6_KRuns great6.9 GB8.1
M5 Max · 36 GBQ6_KRuns great6.9 GB12
M5 Max · 48 GBQ6_KRuns great6.9 GB16
M5 Max · 64 GBQ6_KRuns great6.9 GB16
M5 Max · 128 GBQ6_KRuns great6.9 GB16
M5 Ultra · 96 GBQ6_KRuns great6.9 GB31
M5 Ultra · 256 GBQ6_KRuns great6.9 GB31
M5 Ultra · 512 GBQ6_KRuns great6.9 GB31
M4 · 16 GBQ6_KRuns great6.9 GB3.1
M4 · 24 GBQ6_KRuns great6.9 GB3.1
M4 · 32 GBQ6_KRuns great6.9 GB3.1
M4 Pro · 24 GBQ6_KRuns great6.9 GB6.9
M4 Pro · 48 GBQ6_KRuns great6.9 GB6.9
M4 Pro · 64 GBQ6_KRuns great6.9 GB6.9
M4 Max · 36 GBQ6_KRuns great6.9 GB10
M4 Max · 48 GBQ6_KRuns great6.9 GB13
M4 Max · 64 GBQ6_KRuns great6.9 GB13
M4 Max · 96 GBQ6_KRuns great6.9 GB10
M4 Max · 128 GBQ6_KRuns great6.9 GB13
M3 · 8 GBQ4_K_SRuns well5.2 GB2.8
M3 · 16 GBQ6_KRuns great6.9 GB2.6
M3 · 24 GBQ6_KRuns great6.9 GB2.6
M3 Pro · 18 GBQ6_KRuns great6.9 GB3.9
M3 Pro · 36 GBQ6_KRuns great6.9 GB3.9
M3 Max · 36 GBQ6_KRuns great6.9 GB7.6
M3 Max · 48 GBQ6_KRuns great6.9 GB10
M3 Max · 64 GBQ6_KRuns great6.9 GB10
M3 Max · 96 GBQ6_KRuns great6.9 GB7.6
M3 Max · 128 GBQ6_KRuns great6.9 GB10
M3 Ultra · 96 GBQ6_KRuns great6.9 GB19
M3 Ultra · 256 GBQ6_KRuns great6.9 GB19
M3 Ultra · 512 GBQ6_KRuns great6.9 GB19
M2 · 8 GBQ4_K_SRuns well5.2 GB2.8
M2 · 16 GBQ6_KRuns great6.9 GB2.6
M2 · 24 GBQ6_KRuns great6.9 GB2.6
M2 Pro · 16 GBQ6_KRuns great6.9 GB5.2
M2 Pro · 32 GBQ6_KRuns great6.9 GB5.2
M2 Max · 32 GBQ6_KRuns great6.9 GB10
M2 Max · 64 GBQ6_KRuns great6.9 GB10
M2 Max · 96 GBQ6_KRuns great6.9 GB10
M2 Ultra · 64 GBQ6_KRuns great6.9 GB19
M2 Ultra · 128 GBQ6_KRuns great6.9 GB19
M2 Ultra · 192 GBQ6_KRuns great6.9 GB19
M1 · 8 GBQ4_K_SRuns well5.2 GB1.9
M1 · 16 GBQ6_KRuns great6.9 GB1.8
M1 Pro · 16 GBQ6_KRuns great6.9 GB5
M1 Pro · 32 GBQ6_KRuns great6.9 GB5
M1 Max · 32 GBQ6_KRuns great6.9 GB9.6
M1 Max · 64 GBQ6_KRuns great6.9 GB9.6
M1 Ultra · 64 GBQ6_KRuns great6.9 GB17
M1 Ultra · 128 GBQ6_KRuns great6.9 GB17

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

Information

Maker
EleutherAI
Released
February 2025
Parameters
6.7B
Longest chat
4K tokens
Reads pictures
No
Licence
llama2
Downloads this month
2,796
Sizes
24, 1.5 to 5.5 GB
Kind
Maker's release

Badge for the model card

Shows on Llemma 7B'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