---
license: other
base_model:
- AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4
language: - en
- zh
library_name: llama.cpp
tags: - gguf
- qwen3.6
- nvfp4
- llama.cpp
- multimodal
- vision
- rtx-5090
AEON Qwen3.6 27B Ultimate Uncensored NVFP4 GGUF
This repository contains a GGUF conversion of:
AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4
Files
AEON-Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4.gguf- Main GGUF model
- Converted from the NVFP4 compressed-tensors checkpoint
- Size: ~25GB
mmproj-BF16.gguf- Multimodal / vision projector
- Required only for image input
- Text-only usage does not require this file
Example llama.cpp usage
Text-only:
.\llama-server.exe `
-m "AEON-Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4.gguf" `
--host 0.0.0.0 `
--port 10000 `
-ngl 999 `
-c 32768 `
--flash-attn on
Multimodal:
.\llama-server.exe `
-m "AEON-Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4.gguf" `
--mmproj "mmproj-BF16.gguf" `
--host 0.0.0.0 `
--port 10000 `
-ngl 999 `
-c 32768 `
--flash-attn on
Notes
This GGUF was converted locally on Windows 11 Pro using a patched llama.cpp conversion workflow for NVFP4 compressed-tensors.
The main model was tested locally with llama.cpp-compatible runtime on RTX 5090.
Attribution
Original model:
AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4
Please check and follow the upstream model license and usage terms.