Models

Gemma 3 1B

Google · March 2025

Get

Hugging Face

A 1B chat model from Google.

Downloads: 620K
Likes: 230
Parameters: 1B
Context: 32K
Licence: gemma

Runs great on M4 · 16 GB

The Q5_K_S size from unsloth is the one to get: a 0.8 GB download that uses 1.0 GB of memory, and writes about 51 words a second.

Gemma 3 1B 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 unsloth/gemma-3-1b-it-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.

Size	Download	Memory at 8K	Get
UD-IQ1_S	0.6 GB	0.7 GBestimated	Open
UD-IQ1_M	0.6 GB	0.7 GBestimated	Open
UD-IQ2_XXS	0.6 GB	0.7 GBestimated	Open
UD-IQ2_M	0.6 GB	0.7 GBestimated	Open
UD-IQ3_XXS	0.6 GB	0.7 GBestimated	Open
Q3_K_S	0.7 GB	0.8 GBestimated	Open
Q2_K	0.7 GB	0.8 GBestimated	Open
Q2_K_L	0.7 GB	0.8 GBestimated	Open
UD-Q2_K_XL	0.7 GB	0.8 GBestimated	Open
IQ4_XS	0.7 GB	0.8 GBestimated	Open
IQ4_NL	0.7 GB	0.9 GBestimated	Open
Q3_K_M	0.7 GB	0.9 GBestimated	Open
UD-Q3_K_XL	0.7 GB	0.9 GBestimated	Open
Q4_1	0.8 GB	0.9 GBestimated	Open
Q4_K_S	0.8 GB	0.9 GBestimated	Open
Q4_K_M	0.8 GB	0.9 GBestimated	Open
UD-Q4_K_XL	0.8 GB	0.9 GBestimated	Open
Q5_K_S	0.8 GB	1.0 GBestimated	Open

Run It Anywhere

Gemma 3 1B is an open model, so any app that runs GGUF files can run it. These take the Q5_K_S size from unsloth straight from Hugging Face.

LM StudioMac, Windows, LinuxOpen
JanMac, Windows, LinuxOpen
MstyMac, Windows, LinuxOpen
OllamaMac, Windows, Linuxollama run hf.co/unsloth/gemma-3-1b-it-GGUF:Q5_K_S
llama.cppMac, Windows, Linuxllama-server -hf unsloth/gemma-3-1b-it-GGUF:Q5_K_S
Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/unsloth/gemma-3-1b-it-GGUF:Q5_K_S

On a PC Graphics Card

Video memory	Largest size that fits	Download	Memory at 8K
6 GB	Q5_K_S	0.8 GB	1.0 GB
8 GB	Q5_K_S	0.8 GB	1.0 GB
10 GB	Q5_K_S	0.8 GB	1.0 GB
11 GB	Q5_K_S	0.8 GB	1.0 GB
12 GB	Q5_K_S	0.8 GB	1.0 GB
16 GB	Q5_K_S	0.8 GB	1.0 GB
20 GB	Q5_K_S	0.8 GB	1.0 GB
24 GB	Q5_K_S	0.8 GB	1.0 GB
32 GB	Q5_K_S	0.8 GB	1.0 GB
48 GB	Q5_K_S	0.8 GB	1.0 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

Mac	Size	Fit	Memory	Writes
M6 · 16 GB	Q5_K_S	Runs great	1.0 GB	82
M6 · 24 GB	Q5_K_S	Runs great	1.0 GB	82
M6 · 32 GB	Q5_K_S	Runs great	1.0 GB	82
M5 · 16 GB	Q5_K_S	Runs great	1.0 GB	79
M5 · 24 GB	Q5_K_S	Runs great	1.0 GB	79
M5 · 32 GB	Q5_K_S	Runs great	1.0 GB	79
M5 Pro · 24 GB	Q5_K_S	Runs great	1.0 GB	141
M5 Pro · 48 GB	Q5_K_S	Runs great	1.0 GB	141
M5 Pro · 64 GB	Q5_K_S	Runs great	1.0 GB	141
M5 Max · 36 GB	Q5_K_S	Runs great	1.0 GB	190
M5 Max · 48 GB	Q5_K_S	Runs great	1.0 GB	231
M5 Max · 64 GB	Q5_K_S	Runs great	1.0 GB	231
M5 Max · 128 GB	Q5_K_S	Runs great	1.0 GB	231
M5 Ultra · 96 GB	Q5_K_S	Runs great	1.0 GB	338
M5 Ultra · 256 GB	Q5_K_S	Runs great	1.0 GB	338
M5 Ultra · 512 GB	Q5_K_S	Runs great	1.0 GB	338
M4 · 16 GB	Q5_K_S	Runs great	1.0 GB	51
M4 · 24 GB	Q5_K_S	Runs great	1.0 GB	51
M4 · 32 GB	Q5_K_S	Runs great	1.0 GB	51
M4 Pro · 24 GB	Q5_K_S	Runs great	1.0 GB	82
M4 Pro · 48 GB	Q5_K_S	Runs great	1.0 GB	82
M4 Pro · 64 GB	Q5_K_S	Runs great	1.0 GB	82
M4 Max · 36 GB	Q5_K_S	Runs great	1.0 GB	99
M4 Max · 48 GB	Q5_K_S	Runs great	1.0 GB	109
M4 Max · 64 GB	Q5_K_S	Runs great	1.0 GB	109
M4 Max · 96 GB	Q5_K_S	Runs great	1.0 GB	99
M4 Max · 128 GB	Q5_K_S	Runs great	1.0 GB	109
M3 · 8 GB	Q5_K_S	Runs great	1.0 GB	46
M3 · 16 GB	Q5_K_S	Runs great	1.0 GB	46
M3 · 24 GB	Q5_K_S	Runs great	1.0 GB	46
M3 Pro · 18 GB	Q5_K_S	Runs great	1.0 GB	62
M3 Pro · 36 GB	Q5_K_S	Runs great	1.0 GB	62
M3 Max · 36 GB	Q5_K_S	Runs great	1.0 GB	94
M3 Max · 48 GB	Q5_K_S	Runs great	1.0 GB	108
M3 Max · 64 GB	Q5_K_S	Runs great	1.0 GB	108
M3 Max · 96 GB	Q5_K_S	Runs great	1.0 GB	94
M3 Max · 128 GB	Q5_K_S	Runs great	1.0 GB	108
M3 Ultra · 96 GB	Q5_K_S	Runs great	1.0 GB	138
M3 Ultra · 256 GB	Q5_K_S	Runs great	1.0 GB	138
M3 Ultra · 512 GB	Q5_K_S	Runs great	1.0 GB	138
M2 · 8 GB	Q5_K_S	Runs great	1.0 GB	46
M2 · 16 GB	Q5_K_S	Runs great	1.0 GB	46
M2 · 24 GB	Q5_K_S	Runs great	1.0 GB	46
M2 Pro · 16 GB	Q5_K_S	Runs great	1.0 GB	74
M2 Pro · 32 GB	Q5_K_S	Runs great	1.0 GB	74
M2 Max · 32 GB	Q5_K_S	Runs great	1.0 GB	107
M2 Max · 64 GB	Q5_K_S	Runs great	1.0 GB	107
M2 Max · 96 GB	Q5_K_S	Runs great	1.0 GB	107
M2 Ultra · 64 GB	Q5_K_S	Runs great	1.0 GB	136
M2 Ultra · 128 GB	Q5_K_S	Runs great	1.0 GB	136
M2 Ultra · 192 GB	Q5_K_S	Runs great	1.0 GB	136
M1 · 8 GB	Q5_K_S	Runs great	1.0 GB	31
M1 · 16 GB	Q5_K_S	Runs great	1.0 GB	31
M1 Pro · 16 GB	Q5_K_S	Runs great	1.0 GB	59
M1 Pro · 32 GB	Q5_K_S	Runs great	1.0 GB	59
M1 Max · 32 GB	Q5_K_S	Runs great	1.0 GB	78
M1 Max · 64 GB	Q5_K_S	Runs great	1.0 GB	78
M1 Ultra · 64 GB	Q5_K_S	Runs great	1.0 GB	93
M1 Ultra · 128 GB	Q5_K_S	Runs great	1.0 GB	93

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

Packaged by

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

Information

Maker: Google
Released: March 2025
Parameters: 1B
Longest chat: 32K tokens
Reads pictures: No
Licence: gemma
Downloads this month: 619,824
Likes: 230
Sizes: 18, 0.6 to 0.8 GB
Kind: Maker's release
Source: google/gemma-3-1b-it

Badge for the model card

Shows on Gemma 3 1B's Hugging Face card which Macs run it well, linked to this page. It updates every morning.

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.