Hugging FaceA 1.2B chat model from Meta.
- Rating
- 1.5
- Downloads
- 297K
- Likes
- 420
- Parameters
- 1.2B
- Context
- 128K
- Licence
- llama3.2
Runs great on M4 · 16 GB
The Q8_0 size from unsloth is the one to get: a 1.3 GB download that uses 3.9 GB of memory, and writes about 31 words a second.
Ratings
1.5 out of 5, across people's blind votes on its answers and the tests below.
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.
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.
Llama 3.2 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/Llama-3.2-1B-Instruct-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 | 0.4 GB | 0.7 GBestimated | Open |
| UD-IQ1_M | 0.4 GB | 0.7 GBestimated | Open |
| UD-IQ2_XXS | 0.5 GB | 0.7 GBestimated | Open |
| UD-IQ2_M | 0.5 GB | 0.8 GBestimated | Open |
| UD-IQ3_XXS | 0.6 GB | 0.8 GBestimated | Open |
| Q2_K | 0.6 GB | 0.8 GBestimated | Open |
| Q2_K_L | 0.6 GB | 0.8 GBestimated | Open |
| UD-Q2_K_XL | 0.6 GB | 0.9 GBestimated | Open |
| Q3_K_S | 0.6 GB | 0.9 GBestimated | Open |
| Q3_K_M | 0.7 GB | 1.0 GBestimated | Open |
| UD-Q3_K_XL | 0.7 GB | 1.0 GBestimated | Open |
| IQ4_XS | 0.7 GB | 1.0 GBestimated | Open |
| IQ4_NL | 0.8 GB | 1.0 GBestimated | Open |
| Q4_0 | 0.8 GB | 1.0 GBestimated | Open |
| Q4_K_S | 0.8 GB | 1.0 GBestimated | Open |
| Q4_K_M | 0.8 GB | 1.1 GBestimated | Open |
| Q4_1 | 0.8 GB | 1.1 GBestimated | Open |
| UD-Q4_K_XL | 0.8 GB | 1.1 GBestimated | Open |
| Q5_K_S | 0.9 GB | 1.2 GBestimated | Open |
| Q5_K_M | 0.9 GB | 1.2 GBestimated | Open |
| UD-Q5_K_XL | 0.9 GB | 1.2 GBestimated | Open |
| Q6_K | 1.0 GB | 1.3 GBestimated | Open |
| UD-Q6_K_XL | 1.2 GB | 1.5 GBestimated | Open |
| Q8_0 | 1.3 GB | 1.6 GBestimated | Open |
| UD-Q8_K_XL | 1.8 GB | 2.1 GBestimated | Open |
| F16 | 2.5 GB | 2.7 GBestimated | Open |
| BF16 | 2.5 GB | 2.7 GBestimated | Open |
Run It Anywhere
Llama 3.2 1B is an open model, so any app that runs GGUF files can run it. These take the Q8_0 size from unsloth straight from Hugging Face.
- LM StudioMac, Windows, Linux
Open - JanMac, Windows, Linux
Open - MstyMac, Windows, Linux
Open - OllamaMac, Windows, Linux
ollama run hf.co/unsloth/Llama-3.2-1B-Instruct-GGUF:Q8_0 - llama.cppMac, Windows, Linux
llama-server -hf unsloth/Llama-3.2-1B-Instruct-GGUF:Q8_0 - Docker Model RunnerMac, Windows, Linux
docker model run hf.co/unsloth/Llama-3.2-1B-Instruct-GGUF:Q8_0
On a PC Graphics Card
| Video memory | Largest size that fits | Download | Memory at 8K |
|---|---|---|---|
| 6 GB | BF16 | 2.5 GB | 2.7 GB |
| 8 GB | BF16 | 2.5 GB | 2.7 GB |
| 10 GB | BF16 | 2.5 GB | 2.7 GB |
| 11 GB | BF16 | 2.5 GB | 2.7 GB |
| 12 GB | BF16 | 2.5 GB | 2.7 GB |
| 16 GB | BF16 | 2.5 GB | 2.7 GB |
| 20 GB | BF16 | 2.5 GB | 2.7 GB |
| 24 GB | BF16 | 2.5 GB | 2.7 GB |
| 32 GB | BF16 | 2.5 GB | 2.7 GB |
| 48 GB | BF16 | 2.5 GB | 2.7 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 | Q8_0 | Runs great | 3.9 GB | 46 |
| M6 · 24 GB | Q8_0 | Runs great | 3.9 GB | 46 |
| M6 · 32 GB | Q8_0 | Runs great | 3.9 GB | 46 |
| M5 · 16 GB | Q8_0 | Runs great | 3.9 GB | 42 |
| M5 · 24 GB | Q8_0 | Runs great | 3.9 GB | 42 |
| M5 · 32 GB | Q8_0 | Runs great | 3.9 GB | 42 |
| M5 Pro · 24 GB | Q8_0 | Runs great | 3.9 GB | 82 |
| M5 Pro · 48 GB | Q8_0 | Runs great | 3.9 GB | 82 |
| M5 Pro · 64 GB | Q8_0 | Runs great | 3.9 GB | 82 |
| M5 Max · 36 GB | Q8_0 | Runs great | 3.9 GB | 119 |
| M5 Max · 48 GB | Q8_0 | Runs great | 3.9 GB | 153 |
| M5 Max · 64 GB | Q8_0 | Runs great | 3.9 GB | 153 |
| M5 Max · 128 GB | Q8_0 | Runs great | 3.9 GB | 153 |
| M5 Ultra · 96 GB | Q8_0 | Runs great | 3.9 GB | 271 |
| M5 Ultra · 256 GB | Q8_0 | Runs great | 3.9 GB | 271 |
| M5 Ultra · 512 GB | Q8_0 | Runs great | 3.9 GB | 271 |
| M4 · 16 GB | Q8_0 | Runs great | 3.9 GB | 31 |
| M4 · 24 GB | Q8_0 | Runs great | 3.9 GB | 31 |
| M4 · 32 GB | Q8_0 | Runs great | 3.9 GB | 31 |
| M4 Pro · 24 GB | Q8_0 | Runs great | 3.9 GB | 62 |
| M4 Pro · 48 GB | Q8_0 | Runs great | 3.9 GB | 62 |
| M4 Pro · 64 GB | Q8_0 | Runs great | 3.9 GB | 62 |
| M4 Max · 36 GB | Q8_0 | Runs great | 3.9 GB | 84 |
| M4 Max · 48 GB | Q8_0 | Runs great | 3.9 GB | 102 |
| M4 Max · 64 GB | Q8_0 | Runs great | 3.9 GB | 102 |
| M4 Max · 96 GB | Q8_0 | Runs great | 3.9 GB | 84 |
| M4 Max · 128 GB | Q8_0 | Runs great | 3.9 GB | 102 |
| M3 · 8 GB | Q8_0 | Runs great | 2.7 GB | 26 |
| M3 · 16 GB | Q8_0 | Runs great | 3.9 GB | 26 |
| M3 · 24 GB | Q8_0 | Runs great | 3.9 GB | 26 |
| M3 Pro · 18 GB | Q8_0 | Runs great | 3.9 GB | 38 |
| M3 Pro · 36 GB | Q8_0 | Runs great | 3.9 GB | 38 |
| M3 Max · 36 GB | Q8_0 | Runs great | 3.9 GB | 69 |
| M3 Max · 48 GB | Q8_0 | Runs great | 3.9 GB | 86 |
| M3 Max · 64 GB | Q8_0 | Runs great | 3.9 GB | 86 |
| M3 Max · 96 GB | Q8_0 | Runs great | 3.9 GB | 69 |
| M3 Max · 128 GB | Q8_0 | Runs great | 3.9 GB | 86 |
| M3 Ultra · 96 GB | Q8_0 | Runs great | 3.9 GB | 139 |
| M3 Ultra · 256 GB | Q8_0 | Runs great | 3.9 GB | 139 |
| M3 Ultra · 512 GB | Q8_0 | Runs great | 3.9 GB | 139 |
| M2 · 8 GB | Q8_0 | Runs great | 2.7 GB | 26 |
| M2 · 16 GB | Q8_0 | Runs great | 3.9 GB | 26 |
| M2 · 24 GB | Q8_0 | Runs great | 3.9 GB | 26 |
| M2 Pro · 16 GB | Q8_0 | Runs great | 3.9 GB | 49 |
| M2 Pro · 32 GB | Q8_0 | Runs great | 3.9 GB | 49 |
| M2 Max · 32 GB | Q8_0 | Runs great | 3.9 GB | 86 |
| M2 Max · 64 GB | Q8_0 | Runs great | 3.9 GB | 86 |
| M2 Max · 96 GB | Q8_0 | Runs great | 3.9 GB | 86 |
| M2 Ultra · 64 GB | Q8_0 | Runs great | 3.9 GB | 138 |
| M2 Ultra · 128 GB | Q8_0 | Runs great | 3.9 GB | 138 |
| M2 Ultra · 192 GB | Q8_0 | Runs great | 3.9 GB | 138 |
| M1 · 8 GB | Q8_0 | Runs great | 2.7 GB | 18 |
| M1 · 16 GB | Q8_0 | Runs great | 3.9 GB | 18 |
| M1 Pro · 16 GB | Q8_0 | Runs great | 3.9 GB | 45 |
| M1 Pro · 32 GB | Q8_0 | Runs great | 3.9 GB | 45 |
| M1 Max · 32 GB | Q8_0 | Runs great | 3.9 GB | 74 |
| M1 Max · 64 GB | Q8_0 | Runs great | 3.9 GB | 74 |
| M1 Ultra · 64 GB | Q8_0 | Runs great | 3.9 GB | 110 |
| M1 Ultra · 128 GB | Q8_0 | Runs great | 3.9 GB | 110 |
Writes is words a second, estimated from the Mac's memory speed and the model's size; how, and how wrong.
Packaged by
6 repositories on Hugging Face carry this model for llama.cpp. Outmake reads them all as one model.
Information
- Maker
- Meta
- Released
- September 2024
- Parameters
- 1.2B
- Longest chat
- 128K tokens
- Reads pictures
- No
- Licence
- llama3.2
- Downloads this month
- 296,653
- Likes
- 420
- Sizes
- 27, 0.4 to 2.5 GB
- Kind
- Maker's release
Badge for the model card
Shows on Llama 3.2 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.