Muse-Spark-Open-816B-A42B-oQ3e

This model was quantized using oQ (oMLX v0.7.1) mixed-precision quantization with oQe importance-matrix quantization.

Re-uploaded on 2026-10-16 at 08:19 UTC in the layout of mlx-lm 0.32.2, so it loads in both mlx-lm 0.32.2 and oMLX. This version keeps the lm_head in BF16, which mlx-lm 0.32.2 needs for this model, and gives q, k, v and the attention output gate of each layer one bit width, as mlx-lm stores them as one projection.

A mixed-precision MLX checkpoint of meta-models/Muse-Spark-Open-816B-A42B, Meta's 816B total / 42B active Mixture-of-Experts model with a 1M-token context window, built for coding and agentic work. It includes the language model and the vision encoder. The audio encoder is not included.

The effective average is 3.616 bits/weight for the complete checkpoint.

Quantization and Bit Distribution

oQ3e allocates bits by measured layer sensitivity, then rounds each group with an importance matrix (128 samples x 512 tokens) and a weighted least-squares refit of its scale and bias.

Storage format Logical weights Share of total Tensor storage Effective bits/weight
Affine 3-bit, group 64 798.864B 97.195% 325.500 GiB 3.50
Affine 5-bit, group 64 8.338B 1.014% 5.339 GiB 5.50
Affine 4-bit, group 64 6.275B 0.763% 3.287 GiB 4.50
Affine 8-bit, group 64 4.725B 0.575% 4.676 GiB 8.50
BF16 3.585B 0.436% 6.678 GiB 16.00
FP32 0.132B 0.016% 0.492 GiB 32.00
Total 821.920B 100% 345.972 GiB 3.616

Storage Breakdown

Component Logical weights Storage
Language backbone 819.992B 367.628 GB / 342.381 GiB
Vision encoder 1.928B 3.856 GB / 3.591 GiB
Total 821.920B 371.485 GB / 345.972 GiB

On an M3 Ultra (512 GB) the model takes about 372 GB in oMLX, leaving room for the KV cache.

Use with mlx-lm

pip install -U "mlx-lm>=0.32.2"
mlx_lm.generate --model mlx-community/Muse-Spark-Open-816B-A42B-oQ3e \
  --prompt "Explain convolutional networks like I am 5" \
  --max-tokens 5000

Use with oMLX

Download the model in oMLX and chat with it in the built-in chat UI. You can add images and set the reasoning effort from low to xhigh. Reasoning cannot be turned off, and the default (Auto) uses xhigh.

You can also use

License

Muse Model License, following the source model. See LICENSE.md and USE_POLICY.md.

Downloads last month
1,610
Safetensors
Model size
822B params
Tensor type
BF16
ยท
U32
ยท
F32
ยท
MLX
Hardware compatibility
Log In to add your hardware

3-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for mlx-community/Muse-Spark-Open-816B-A42B-oQ3e

Quantized
(25)
this model

Collection including mlx-community/Muse-Spark-Open-816B-A42B-oQ3e