Downloads · 30 days
68
9% of all-time downloads
Qwen3.8-27B-MLX-MXFP4-MTP
Shiftedx/Qwen3.8-27B-MLX-MXFP4-MTP
Qwen3.8-27B-MLX-MXFP4-MTP is a image-text-to-text model from Shiftedx. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for mlx. The card lists the license as apache-2.0.
Vision-enabled MTPLX package of Qwen/Qwen3.8-27B, pinned to revision 1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0.
All-time downloads
737
Public
Parameters
27.4B
16.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors16.1 GB · 100%
Parameter types
How the weights are stored.
U3226.9B · 98%
Try a prompt
Base models
Model card
From the Hugging Face model README
Qwen3.8-27B MLX MXFP4 + Native MTP
Vision-enabled MTPLX package of Qwen/Qwen3.8-27B, pinned to revision
1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0.
- Language model: MXFP4, 4-bit, group size 32
- Vision tower: same-revision BF16 weights
- Speculation: 15 native BF16 MTP tensors; recommended depth 3
- Runtime: MTPLX 2.0.2 or newer
The Hub's approximately 5.5B safetensors count reflects packed MXFP storage elements; the underlying architecture remains the full 27B model.
Serve
mtplx quickstart \
--model Shiftedx/Qwen3.8-27B-MLX-MXFP4-MTP \
--mtp --depth 3 --profile sustained
The MTP tensor gate, text generation, and three image requests through the local OpenAI-compatible API passed. Quantization and speculative decoding can still change behavior, so independently evaluate important use cases. Treat prompts, images, and outputs as untrusted: do not submit secrets, and sandbox tools or generated code with least-privilege access. This package adds no telemetry or remote execution.
The upstream Apache-2.0 license and model limitations continue to apply.
- Task
- Image-Text-to-Text
- Library
- mlx
- Type
- qwen3_5
- License
- apache-2.0
- Created
- Aug 14, 2026
- Updated
- Aug 14, 2026