MLX 4-bit build (4.5 bits per weight, about 14 GB) of Qwen3.8-27B-RFT-v1 for Apple silicon. Converted with mlx-lm 0.32.2.
A model by Horizon Tech d.o.o., Sarajevo, Bosnia and Herzegovina.
| Developed by | Horizon Tech d.o.o., Sarajevo, Bosnia and Herzegovina |
| Model type | MLX 4-bit build of a 27B-parameter language model, text only |
| Converted from | Qwen3.8-27B-RFT-v1 (fine-tune of Qwen/Qwen3.8-27B) |
| Format | MLX, 4-bit (4.5 bits per weight), 14 GB |
| Not included | The vision tower and the MTP head (mlx-lm does not use them) |
| License | Apache 2.0 |
| Released | November 2026 |
| Author | Hamza Abdagic (ORCID 0009-0002-4078-4922) |
| Contact | hamza.abdagic@horizon-tech.io |
| Repository | Format | Size | For |
|---|---|---|---|
| Qwen3.8-27B-RFT-v1 | bfloat16 safetensors | 52 GB | Transformers, vLLM, SGLang; further fine-tuning |
| Qwen3.8-27B-RFT-v1-NVFP4 | NVFP4 (4-bit) | 20 GB | vLLM on NVIDIA GPUs with FP4 support |
| Qwen3.8-27B-RFT-v1-mlx-4bit | MLX 4-bit | 14 GB | Apple silicon, text only |
pip install mlx-lm
mlx_lm.chat --model ghostrider0470/Qwen3.8-27B-RFT-v1-mlx-4bit
@misc{horizontech2026rftv1mlx,
author = {Abdagic, Hamza},
title = {Qwen3.8-27B-RFT-v1-mlx-4bit},
year = {2026},
publisher = {Hugging Face},
organization = {Horizon Tech d.o.o., Sarajevo},
howpublished = {\url{https://huggingface.co/ghostrider0470/Qwen3.8-27B-RFT-v1-mlx-4bit}}
}
Apache 2.0, as the base model.
model-00002-of-00003.safetensors 5.4 GBmodel-00001-of-00003.safetensors 5.3 GBmodel-00003-of-00003.safetensors 4.5 GBghostrider0470/Qwen3.8-27B-RFT-v1