Request access under the MiniMax license

Manual review. Read LICENSE and the territorial, commercial and acceptable-use conditions before requesting access.

MiniMax H3 Image MLX · Q8 affine core

Fifty main DiT blocks and two token-refiner blocks use MLX affine 8-bit weight quantization, group size 64. It reduces storage and memory but changes images; no generation-speed advantage is established.

Community adaptation by Devin Lai. Powered by MiniMax-H3-Image.

Start here: install and quick start · Compare variants · Evidence and limitations

Storage and historical performance · M5 Pro / 64 GiB

Measure	Result
Q8 variant payload	19.83 GiB
One historical warm product trial: denoising	72.28 s
Same trial: peak MLX allocation	21.48 GiB

The variant payload excludes approximately 3.86 GiB of shared components and the roughly 62 GiB text encoder. The historical trial was one 2048 × 2048, six-step product run with cached conditioning and resident weights, using online Q8 on an M5 Pro / 64 GiB. Denoising is not end-to-end time; MLX allocation is not total process or system memory and does not establish minimum RAM. This was not a fresh performance test of the downloadable export. Q8 changes images and has no established speed advantage. Method and full comparisons.

Download only this variant

After installing the main runtime and obtaining the required access:

h3-mlx download --variant q8 --model-dir models/q8
h3-mlx generate --model models/q8 \
  --text-encoder models/q8/text_encoder \
  'An editorial photograph of a cobalt perfume bottle on pale stone' \
  --seed 731 --output outputs/bottle.png

This fetches this variant, required shared components and the separately pinned native text encoder (roughly 62 GiB). It does not fetch other variants. Use --text-encoder-path /path/to/text_encoder to reuse an existing encoder. All image variants retain the six original PDD heads and native FP32 islands. The published compact format requires this runtime, not a generic Diffusers or mlx-lm loader. Tested on M5 Pro / 64 GiB; smaller Macs are unqualified.

Model card and terms

This is an inference port of MiniMaxAI/MiniMax-H3-Image at revision 6534ea1c979a5f8cb00c2bd1025d5ca4ee8d8901, not a newly trained model. No new training dataset is introduced. Training provenance, potential biases, intended uses and upstream limitations remain those of the original model. See modifications, license and NOTICE.

The MiniMax-H3-Image Community License Agreement and its acceptable-use policy apply, including territorial and commercial restrictions. The standard grant excludes the EU, UK, Republic of Korea and US. Access is subject to manual review; it does not provide separate permission. The upstream text encoder retains its own MiniMax H3 agreement. This is not an unrestricted open-source license or an official MiniMax endorsement.

Release: v0.4.1 (v0.4.0 weights with download-count metadata). Access requests are reviewed manually under the retained license. The actual file manifest records exact sizes and SHA-256 hashes.

Report reproducible results or issues · Follow devin-lai