debais/G4-Moonlight-Dusk-26B-A4B-APEX-GGUF overview
G4 Moonlight Dusk 26B A4B APEX GGUF APEX Adaptive Precision for EXpert Models quantizations of Vortex5/G4 Moonlight Dusk 26B A4B https://huggingface.co/Vortex5…
Runs locally from ~11.73 GB disk (12 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| G4-Moonlight-Dusk-APEX-Balanced.gguf | GGUF | GGUF | 20.17 GB | Download |
| G4-Moonlight-Dusk-APEX-Compact.gguf | GGUF | GGUF | 14.18 GB | Download |
| G4-Moonlight-Dusk-APEX-I-Balanced.gguf | GGUF | GGUF | 20.17 GB | Download |
| G4-Moonlight-Dusk-APEX-I-Compact.gguf | GGUF | GGUF | 14.18 GB | Download |
| G4-Moonlight-Dusk-APEX-I-Mini.gguf | GGUF | GGUF | 11.73 GB | Download |
| G4-Moonlight-Dusk-APEX-I-Quality.gguf | GGUF | GGUF | 19.73 GB | Download |
| G4-Moonlight-Dusk-APEX-Quality.gguf | GGUF | GGUF | 19.73 GB | Download |
| G4-Moonlight-Dusk-f16.gguf | GGUF | F16 | 48.41 GB | Download |
Model Details
Model README
---
license: gemma
base_model:
- Vortex5/G4-Moonlight-Dusk-26B-A4B
---
G4-Moonlight-Dusk-26B-A4B-APEX-GGUF
APEX (Adaptive Precision for EXpert Models) quantizations of Vortex5/G4-Moonlight-Dusk-26B-A4B.
I used the script from LocalAI team | APEX Project | Technical Report
I also used the latest chat_template.jinja(2026-07-22 , if google ever update the template again , override the template with the latest) from Google , which fixes a lot of issue , for RP specifically thinking will not longer leak out .
imatrix data is from mradermacher
Run debais/G4-Moonlight-Dusk-26B-A4B-APEX-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models