Hugging FaceA 426B, 26B active chat model from MiniMax that reads pictures as well as text.
- Rating
- 4.7
- Downloads
- 14K
- Likes
- 180
- Parameters
- 426B, 26B active
- Context
- 1M
- Licence
- other
Too big for this Mac
Its smallest size needs 165 GB of memory, more than M4 · 16 GB can give it.
Ratings
4.7 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.
LiveBench: reasoning, coding, maths, data and language on fresh questions. SWE-bench Verified: real GitHub issues fixed so the project's tests pass. 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.
MiniMax M3 runs well from a Mac with an M1 Ultra and 128 GB of memory. 10 of 59 Mac configurations run it well.
Compare with
The nearest models in size.
Sizes
From unsloth/MiniMax-M3-GGUF on Hugging Face. A smaller size is a smaller download and a little less accurate.
| Size | Download | Memory at 8K | Get |
|---|---|---|---|
| UD-IQ1_M | 129 GB | 130 GBestimated | Open |
| UD-IQ2_XXS | 135 GB | 136 GBestimated | Open |
| UD-IQ2_M | 135 GB | 136 GBestimated | Open |
| UD-Q2_K_XL | 144 GB | 145 GBestimated | Open |
| UD-IQ3_XXS | 160 GB | 161 GBestimated | Open |
| UD-IQ3_S | 176 GB | 176 GBestimated | Open |
| UD-Q3_K_M | 195 GB | 196 GBestimated | Open |
| UD-Q3_K_XL | 196 GB | 196 GBestimated | Open |
| UD-IQ4_XS | 208 GB | 209 GBestimated | Open |
| UD-IQ4_NL | 213 GB | 213 GBestimated | Open |
| UD-Q4_K_S | 249 GB | 249 GBestimated | Open |
| MXFP4_MOE | 257 GB | 258 GBestimated | Open |
| UD-Q4_K_M | 265 GB | 266 GBestimated | Open |
| UD-Q4_K_XL | 266 GB | 267 GBestimated | Open |
| UD-Q5_K_S | 300 GB | 300 GBestimated | Open |
| UD-Q5_K_M | 319 GB | 319 GBestimated | Open |
| UD-Q5_K_XL | 319 GB | 320 GBestimated | Open |
| UD-Q6_K | 355 GB | 356 GBestimated | Open |
| UD-Q6_K_XL | 388 GB | 389 GBestimated | Open |
| Q8_0 | 454 GB | 454 GBestimated | Open |
| UD-Q8_K_XL | 465 GB | 466 GBestimated | Open |
| BF16 | 853 GB | 854 GBestimated | Open |
Run It Anywhere
MiniMax M3 is an open model, so any app that runs GGUF files can run it. These take the UD-IQ1_M 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/MiniMax-M3-GGUF:UD-IQ1_M - llama.cppMac, Windows, Linux
llama-server -hf unsloth/MiniMax-M3-GGUF:UD-IQ1_M - Docker Model RunnerMac, Windows, Linux
docker model run hf.co/unsloth/MiniMax-M3-GGUF:UD-IQ1_M
On a PC Graphics Card
| Video memory | Largest size that fits | Download | Memory at 8K |
|---|---|---|---|
| 6 GB | None worth running | needs 136 GB | |
| 8 GB | None worth running | needs 136 GB | |
| 10 GB | None worth running | needs 136 GB | |
| 11 GB | None worth running | needs 136 GB | |
| 12 GB | None worth running | needs 136 GB | |
| 16 GB | None worth running | needs 136 GB | |
| 20 GB | None worth running | needs 136 GB | |
| 24 GB | None worth running | needs 136 GB | |
| 32 GB | None worth running | needs 136 GB | |
| 48 GB | None worth running | needs 136 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 | IQ1_S | Too big | 165 GB | |
| M6 · 24 GB | IQ1_S | Too big | 165 GB | |
| M6 · 32 GB | IQ1_S | Too big | 165 GB | |
| M5 · 16 GB | IQ1_S | Too big | 165 GB | |
| M5 · 24 GB | IQ1_S | Too big | 165 GB | |
| M5 · 32 GB | IQ1_S | Too big | 165 GB | |
| M5 Pro · 24 GB | IQ1_S | Too big | 165 GB | |
| M5 Pro · 48 GB | IQ1_S | Too big | 165 GB | |
| M5 Pro · 64 GB | IQ1_S | Too big | 165 GB | |
| M5 Max · 36 GB | IQ1_S | Too big | 165 GB | |
| M5 Max · 48 GB | IQ1_S | Too big | 165 GB | |
| M5 Max · 64 GB | IQ1_S | Too big | 165 GB | |
| M5 Max · 128 GB | IQ1_S | Runs well | 100 GB | 23 |
| M5 Ultra · 96 GB | IQ1_S | Too big | 165 GB | |
| M5 Ultra · 256 GB | UD-IQ3_XXS | Runs great | 170 GB | 30 |
| M5 Ultra · 512 GB | UD-Q5_K_XL | Runs great | 338 GB | 19 |
| M4 · 16 GB | IQ1_S | Too big | 165 GB | |
| M4 · 24 GB | IQ1_S | Too big | 165 GB | |
| M4 · 32 GB | IQ1_S | Too big | 165 GB | |
| M4 Pro · 24 GB | IQ1_S | Too big | 165 GB | |
| M4 Pro · 48 GB | IQ1_S | Too big | 165 GB | |
| M4 Pro · 64 GB | IQ1_S | Too big | 165 GB | |
| M4 Max · 36 GB | IQ1_S | Too big | 165 GB | |
| M4 Max · 48 GB | IQ1_S | Too big | 165 GB | |
| M4 Max · 64 GB | IQ1_S | Too big | 165 GB | |
| M4 Max · 96 GB | IQ1_S | Too big | 165 GB | |
| M4 Max · 128 GB | IQ1_S | Runs well | 100 GB | 16 |
| M3 · 8 GB | IQ1_S | Too big | 165 GB | |
| M3 · 16 GB | IQ1_S | Too big | 165 GB | |
| M3 · 24 GB | IQ1_S | Too big | 165 GB | |
| M3 Pro · 18 GB | IQ1_S | Too big | 165 GB | |
| M3 Pro · 36 GB | IQ1_S | Too big | 165 GB | |
| M3 Max · 36 GB | IQ1_S | Too big | 165 GB | |
| M3 Max · 48 GB | IQ1_S | Too big | 165 GB | |
| M3 Max · 64 GB | IQ1_S | Too big | 165 GB | |
| M3 Max · 96 GB | IQ1_S | Too big | 165 GB | |
| M3 Max · 128 GB | IQ1_S | Runs well | 100 GB | 14 |
| M3 Ultra · 96 GB | IQ1_S | Too big | 165 GB | |
| M3 Ultra · 256 GB | UD-IQ3_XXS | Runs great | 170 GB | 17 |
| M3 Ultra · 512 GB | UD-Q5_K_XL | Runs great | 338 GB | 11 |
| M2 · 8 GB | IQ1_S | Too big | 165 GB | |
| M2 · 16 GB | IQ1_S | Too big | 165 GB | |
| M2 · 24 GB | IQ1_S | Too big | 165 GB | |
| M2 Pro · 16 GB | IQ1_S | Too big | 165 GB | |
| M2 Pro · 32 GB | IQ1_S | Too big | 165 GB | |
| M2 Max · 32 GB | IQ1_S | Too big | 165 GB | |
| M2 Max · 64 GB | IQ1_S | Too big | 165 GB | |
| M2 Max · 96 GB | IQ1_S | Too big | 165 GB | |
| M2 Ultra · 64 GB | IQ1_S | Too big | 165 GB | |
| M2 Ultra · 128 GB | IQ1_S | Runs well | 100 GB | 22 |
| M2 Ultra · 192 GB | UD-IQ1_M | Runs great | 131 GB | 19 |
| M1 · 8 GB | IQ1_S | Too big | 165 GB | |
| M1 · 16 GB | IQ1_S | Too big | 165 GB | |
| M1 Pro · 16 GB | IQ1_S | Too big | 165 GB | |
| M1 Pro · 32 GB | IQ1_S | Too big | 165 GB | |
| M1 Max · 32 GB | IQ1_S | Too big | 165 GB | |
| M1 Max · 64 GB | IQ1_S | Too big | 165 GB | |
| M1 Ultra · 64 GB | IQ1_S | Too big | 165 GB | |
| M1 Ultra · 128 GB | IQ1_S | Runs well | 100 GB | 17 |
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
- MiniMax
- Released
- June 2026
- Parameters
- 426B, 26B active
- Longest chat
- 1M tokens
- Reads pictures
- Yes
- Licence
- other
- Downloads this month
- 13,838
- Likes
- 180
- Sizes
- 22, 129 to 853 GB
- Kind
- Maker's release
- Source
- MiniMaxAI/MiniMax-M3
Badge for the model card
Shows on MiniMax M3'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.