You can deploy this model and start earning money today!

majentik/
Mistral-Medium-3.5-128B-MLX-6bit

Name: Mistral-Medium-3.5-128B-MLX-6bit
Availability: InStock

mistralai/Mistral-Medium-3.5-128B

quantized

other

6-bit

Views

No views yet

Mistral-Medium-3.5-128B-MLX-6bit

MLX 6bit (affine, group size 64) quantized variant of mistralai/Mistral-Medium-3.5-128B (text tower quantized; vision tower and projector retained in BF16; source checkpoint is FP8) for Apple silicon via mlx-lm.

Provenance

Source: mistralai/Mistral-Medium-3.5-128B @ revision 22b2b868a15677cfa6061277ed2f653d1349a9ab (Modified MIT — revenue carve-out below $20M/month).
Quantized with mlx_lm.convert (mlx-lm 0.31.3): affine, 6-bit, group size 64.

Smoke gate

Before upload this pack passed a deterministic coherence gate: greedy 48-token chat generation loaded through mlx_lm.load, judged for emptiness, repetition loops, multi-script gibberish, and special-token debris. Verdict: ok.

Usage

bash

pip install mlx-lm
mlx_lm.generate --model majentik/Mistral-Medium-3.5-128B-MLX-6bit --prompt "Hello"

License

Modified MIT (upstream LICENSE): free use with a revenue carve-out — organisations above $20M/month revenue need a commercial agreement with Mistral AI. This is not a plain MIT license; read the upstream terms before commercial deployment.

Evaluation

Benchmark	Score
arc_easy_acc	0.7800
hellaswag_acc	0.5250

Available tiers

Repo	Quant
majentik/Mistral-Medium-3.5-128B-MLX-2bit	2bit
majentik/Mistral-Medium-3.5-128B-MLX-3bit	3bit
majentik/Mistral-Medium-3.5-128B-MLX-4bit	4bit
majentik/Mistral-Medium-3.5-128B-MLX-5bit	5bit
majentik/Mistral-Medium-3.5-128B-MLX-6bit	6bit
majentik/Mistral-Medium-3.5-128B-MLX-MXFP4	MXFP4