MuXodious GLM 4.7 Flash absolute heresy APEX needs about 20.4 GB of memory at Q4_K_M (a 19.1 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 4 quants, from 11.3 GB (IQ3_XXS) to 21.7 GB (Q5_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 3.4 GB.
MuXodious GLM 4.7 Flash absolute heresy APEX – 30B MoE model
MuXodious GLM 4.7 Flash absolute heresy APEX is a 30B‑A3B MoE model based on the GLM family and deepseek2 architecture. It is fine‑tuned with the Heretic abliteration engine and supports roleplay tasks. The model is released under the MIT license.
Good at
- Roleplay
- Large context handling
- MoE efficiency
Limits
- No vision support
- Requires 8k context for memory estimates
- No explicit language support listed
Best for
- Chatbot roleplay scenarios
- Agentic multi‑turn tasks
- Local deployment with llama.cpp
Who it’s for. Developers and researchers needing a 30B MoE model for roleplay and agentic tasks. The model is suitable for local inference on systems with sufficient VRAM and RAM.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ3_XXS | 11.3 GB | 12.7 GB | Too big | Get file → | |
| Q3_K_M | 14.1 GB | 15.5 GB | Too big | Get file → | |
| Q4_K_M | 19.1 GB | 20.4 GB | Too big | Get file → | |
| Q5_K_M | 21.7 GB | 23.0 GB | Too big | Get file → |
For 8 GB VRAM, use IQ3_XXS (file size 11.34 GB, estimated memory 12.7 GB). For 12 GB VRAM, Q3_K_M (14.12 GB, 15.5 GB). For 16 GB VRAM, Q4_K_M (19.06 GB, 20.4 GB). For 20 GB VRAM, Q5_K_M (21.7 GB, 23 GB). Estimates are approximate.
What GPU do you need for MuXodious GLM 4.7 Flash absolute heresy APEX?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: the best quant that fits is Q3_K_M (14.1 GB file, about 15.5 GB in use).
- 24 GB VRAM: the best quant that fits is Q5_K_M (21.7 GB file, about 23.0 GB in use).
Which MuXodious GLM 4.7 Flash absolute heresy APEX quant should you choose?
Q4_K_M is the usual starting point for MuXodious GLM 4.7 Flash absolute heresy APEX: it balances file size and quality well. Here it is a 19.1 GB file.
If memory is tight, the smallest file is IQ3_XXS at 11.3 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about MuXodious GLM 4.7 Flash absolute heresy APEX
How much VRAM does MuXodious GLM 4.7 Flash absolute heresy APEX need?
About 20.4 GB at Q4_K_M with an 8K context window. That is the 19.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run MuXodious GLM 4.7 Flash absolute heresy APEX on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 12.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of MuXodious GLM 4.7 Flash absolute heresy APEX GGUF?
It depends on the quant: IQ3_XXS is 11.3 GB, Q3_K_M is 14.1 GB, Q4_K_M is 19.1 GB, Q5_K_M is 21.7 GB.
What context length does MuXodious GLM 4.7 Flash absolute heresy APEX support?
The GGUF header lists a trained context length of 202,752 tokens (198K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.4 GB at 32K.
Where can I download MuXodious GLM 4.7 Flash absolute heresy APEX GGUF?
The files are on Hugging Face in the Koshkasa/MuXodious_GLM-4.7-Flash-absolute-heresy-APEX-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does MuXodious GLM 4.7 Flash absolute heresy APEX use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run this model on 8 GB VRAM?
Use the IQ3_XXS quantization; it requires about 12.7 GB of VRAM at 8k context.
Does this model support vision tasks?
No, the model does not include vision capabilities.
What license applies to this model?
The model is released under the MIT license.