Models

Muse Spark Open 816B A42B

meta-models · September 2026

Get

Hugging Face

A 820B, 16B active chat model from meta-models that reads pictures as well as text.

Rating: 4.9
Downloads: 21K
Likes: 85
Parameters: 820B, 16B active
Context: 1M
Licence: other

Too big for this Mac

Its smallest size needs 216 GB of memory, more than M4 · 16 GB can give it.

Ratings

4.9 out of 5, across the tests below.

ScienceGPQA Diamond: 89%
Expert questionsHumanity's Last Exam: 61%

GPQA Diamond: questions PhD scientists write for each other. Humanity's Last Exam: questions written to be beyond today's models. 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.

Muse Spark Open 816B A42B runs well from a Mac with an M3 Ultra and 256 GB of memory. 4 of 59 Mac configurations run it well.

Compare with

The nearest models in size.

Sizes

From unsloth/Muse-Spark-Open-816B-A42B-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	176 GB	176 GBestimated	Open
UD-IQ1_M	209 GB	210 GBestimated	Open
UD-IQ2_XXS	226 GB	226 GBestimated	Open
UD-IQ2_S	275 GB	276 GBestimated	Open
UD-Q2_K_XL	311 GB	312 GBestimated	Open
UD-IQ3_XXS	340 GB	341 GBestimated	Open
UD-Q3_K_XL	401 GB	402 GBestimated	Open
UD-Q4_K_XL	506 GB	506 GBestimated	Open

Run It Anywhere

Muse Spark Open 816B A42B is an open model, so any app that runs GGUF files can run it. These take the UD-IQ1_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/Muse-Spark-Open-816B-A42B-GGUF:UD-IQ1_S
llama.cppMac, Windows, Linuxllama-server -hf unsloth/Muse-Spark-Open-816B-A42B-GGUF:UD-IQ1_S
Docker Model RunnerMac, Windows, Linuxdocker model run hf.co/unsloth/Muse-Spark-Open-816B-A42B-GGUF:UD-IQ1_S

On a PC Graphics Card

Video memory	Largest size that fits	Memory at 8K
6 GB	None worth running	needs 312 GB
8 GB	None worth running	needs 312 GB
10 GB	None worth running	needs 312 GB
11 GB	None worth running	needs 312 GB
12 GB	None worth running	needs 312 GB
16 GB	None worth running	needs 312 GB
20 GB	None worth running	needs 312 GB
24 GB	None worth running	needs 312 GB
32 GB	None worth running	needs 312 GB
48 GB	None worth running	needs 312 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	UD-IQ1_S	Too big	216 GB
M6 · 24 GB	UD-IQ1_S	Too big	216 GB
M6 · 32 GB	UD-IQ1_S	Too big	216 GB
M5 · 16 GB	UD-IQ1_S	Too big	216 GB
M5 · 24 GB	UD-IQ1_S	Too big	216 GB
M5 · 32 GB	UD-IQ1_S	Too big	216 GB
M5 Pro · 24 GB	UD-IQ1_S	Too big	216 GB
M5 Pro · 48 GB	UD-IQ1_S	Too big	216 GB
M5 Pro · 64 GB	UD-IQ1_S	Too big	216 GB
M5 Max · 36 GB	UD-IQ1_S	Too big	216 GB
M5 Max · 48 GB	UD-IQ1_S	Too big	216 GB
M5 Max · 64 GB	UD-IQ1_S	Too big	216 GB
M5 Max · 128 GB	UD-IQ1_S	Too big	216 GB
M5 Ultra · 96 GB	UD-IQ1_S	Too big	216 GB
M5 Ultra · 256 GB	UD-IQ1_S	Runs well	196 GB	60
M5 Ultra · 512 GB	UD-Q3_K_XL	Runs well	407 GB	39
M4 · 16 GB	UD-IQ1_S	Too big	216 GB
M4 · 24 GB	UD-IQ1_S	Too big	216 GB
M4 · 32 GB	UD-IQ1_S	Too big	216 GB
M4 Pro · 24 GB	UD-IQ1_S	Too big	216 GB
M4 Pro · 48 GB	UD-IQ1_S	Too big	216 GB
M4 Pro · 64 GB	UD-IQ1_S	Too big	216 GB
M4 Max · 36 GB	UD-IQ1_S	Too big	216 GB
M4 Max · 48 GB	UD-IQ1_S	Too big	216 GB
M4 Max · 64 GB	UD-IQ1_S	Too big	216 GB
M4 Max · 96 GB	UD-IQ1_S	Too big	216 GB
M4 Max · 128 GB	UD-IQ1_S	Too big	216 GB
M3 · 8 GB	UD-IQ1_S	Too big	216 GB
M3 · 16 GB	UD-IQ1_S	Too big	216 GB
M3 · 24 GB	UD-IQ1_S	Too big	216 GB
M3 Pro · 18 GB	UD-IQ1_S	Too big	216 GB
M3 Pro · 36 GB	UD-IQ1_S	Too big	216 GB
M3 Max · 36 GB	UD-IQ1_S	Too big	216 GB
M3 Max · 48 GB	UD-IQ1_S	Too big	216 GB
M3 Max · 64 GB	UD-IQ1_S	Too big	216 GB
M3 Max · 96 GB	UD-IQ1_S	Too big	216 GB
M3 Max · 128 GB	UD-IQ1_S	Too big	216 GB
M3 Ultra · 96 GB	UD-IQ1_S	Too big	216 GB
M3 Ultra · 256 GB	UD-IQ1_S	Runs well	196 GB	29
M3 Ultra · 512 GB	UD-Q3_K_XL	Runs well	407 GB	21
M2 · 8 GB	UD-IQ1_S	Too big	216 GB
M2 · 16 GB	UD-IQ1_S	Too big	216 GB
M2 · 24 GB	UD-IQ1_S	Too big	216 GB
M2 Pro · 16 GB	UD-IQ1_S	Too big	216 GB
M2 Pro · 32 GB	UD-IQ1_S	Too big	216 GB
M2 Max · 32 GB	UD-IQ1_S	Too big	216 GB
M2 Max · 64 GB	UD-IQ1_S	Too big	216 GB
M2 Max · 96 GB	UD-IQ1_S	Too big	216 GB
M2 Ultra · 64 GB	UD-IQ1_S	Too big	216 GB
M2 Ultra · 128 GB	UD-IQ1_S	Too big	216 GB
M2 Ultra · 192 GB	UD-IQ1_S	Too big	216 GB
M1 · 8 GB	UD-IQ1_S	Too big	216 GB
M1 · 16 GB	UD-IQ1_S	Too big	216 GB
M1 Pro · 16 GB	UD-IQ1_S	Too big	216 GB
M1 Pro · 32 GB	UD-IQ1_S	Too big	216 GB
M1 Max · 32 GB	UD-IQ1_S	Too big	216 GB
M1 Max · 64 GB	UD-IQ1_S	Too big	216 GB
M1 Ultra · 64 GB	UD-IQ1_S	Too big	216 GB
M1 Ultra · 128 GB	UD-IQ1_S	Too big	216 GB

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

Packaged by

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

Information

Maker: meta-models
Released: September 2026
Parameters: 820B, 16B active
Longest chat: 1M tokens
Reads pictures: Yes
Licence: other
Downloads this month: 20,979
Likes: 85
Sizes: 8, 176 to 506 GB
Kind: Maker's release
Source: meta-models/Muse-Spark-Open-816B-A42B

Badge for the model card

Shows on Muse Spark Open 816B A42B'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.