---
base_model:
- ubergarm/Qwen3.6-27B-GGUF
---
[Qwen3.6-27B-smol-MTP-IQ4_NL](https://huggingface.co/ubergarm/Qwen3.6-27B-GGUF/blob/main/Qwen3.6-27B-smol-IQ4_NL.gguf) from [ubergarm](https://huggingface.co/ubergarm) with grafted [MTP/draft](https://huggingface.co/IHaveNoClueAndIMustPost/Qwen3.6-27b-MTP-TENSORS-ONLY) tensors. Optimized for older hardware and/or Vulkan acceleration. Fits 131K context on 24GB depending on cache quantization levels and whether vision is loaded on GPU or not.