GLM 4.7 Flash heretic needs about 19.5 GB of memory at Q4_K_M (a 18.1 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 3 quants, from 18.1 GB (Q4_K_M) to 31.8 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 3.4 GB.
GLM 4.7 Flash heretic – 30B MoE model for lightweight deployment
GLM 4.7 Flash heretic is a 30B‑A3B MoE model derived from GLM‑4.7‑Flash. It offers a balance of performance and efficiency for local inference. The model is available in multiple GGUF quantizations for llama.cpp.
Good at
- Large context handling
- MoE efficiency
- Abilitation removal
Limits
- No vision support
- License: other
- Requires 8k context for memory estimates
Best for
- Chat applications
- Long‑form content generation
- Agentic tasks
Who it’s for. Developers and researchers needing a 30B model that can run on consumer GPUs with reduced memory. The model is suitable for local deployment via llama.cpp, vLLM, or SGLang.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 18.1 GB | 19.5 GB | Too big | Get file → | |
| Q6_K | 24.6 GB | 26.0 GB | Too big | Get file → | |
| Q8_0 | 31.8 GB | 33.2 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (file size 18.13 GB, estimated memory 19.5 GB). For 12 GB VRAM, Q6_K (24.61 GB, 26 GB). For 16 GB VRAM, Q8_0 (31.84 GB, 33.2 GB). Estimates are approximate.
What GPU do you need for GLM 4.7 Flash heretic?
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: no quant fits at an 8K context.
- 24 GB VRAM: the best quant that fits is Q4_K_M (18.1 GB file, about 19.5 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (24.6 GB file, about 26.0 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (31.8 GB file, about 33.2 GB in use).
Which GLM 4.7 Flash heretic quant should you choose?
Q4_K_M is the usual starting point for GLM 4.7 Flash heretic: it balances file size and quality well. Here it is a 18.1 GB file.
The Q8_0 file is 31.8 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about GLM 4.7 Flash heretic
How much VRAM does GLM 4.7 Flash heretic need?
About 19.5 GB at Q4_K_M with an 8K context window. That is the 18.1 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 GLM 4.7 Flash heretic on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 19.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of GLM 4.7 Flash heretic GGUF?
It depends on the quant: Q4_K_M is 18.1 GB, Q6_K is 24.6 GB, Q8_0 is 31.8 GB.
What context length does GLM 4.7 Flash heretic 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 GLM 4.7 Flash heretic GGUF?
The files are on Hugging Face in the ThalisAI/GLM-4.7-Flash-heretic repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does GLM 4.7 Flash heretic use?
Hugging Face lists the license as other. 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 GLM 4.7 Flash heretic on 8 GB VRAM?
Use the Q4_K_M quantization; it requires about 19.5 GB of VRAM at 8k context, which is above 8 GB, so it is not feasible.
Does this model support vision tasks?
No, the model has no vision capabilities.
What license applies to this model?
The model is released under an "other" license; check the repository for details.