Qwen3.8 Flash Next Heretic needs about 57.4 GB of memory at Q5_K_M (a 56.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 2 quants, from 56.1 GB (Q5_K_M) to 93.6 GB (IQ4_XS). Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.
Qwen3.8 Flash Next Heretic – 170B sparse MoE model with vision
Qwen3.8 Flash Next Heretic is a 170B parameter sparse MoE model based on qwen4exp architecture. It supports vision via F16 mmproj and offers a native context of 262,144 tokens. The model is uncensored and includes 512 experts with 10 active per token.
Good at
- Large context handling
- Sparse MoE efficiency
- Vision input support
Limits
- No explicit license details
- Requires significant VRAM for full performance
- No text-only restriction
Best for
- Long‑form content generation
- Image‑text multimodal tasks
- Research on sparse MoE models
Who it’s for. Developers and researchers needing a large‑scale, vision‑capable LLM with sparse MoE architecture. Suitable for experimentation and deployment on systems with adequate VRAM.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q5_K_M | 56.1 GB (32 parts) | 57.4 GB | Too big | Get file → | |
| IQ4_XS | 93.6 GB (41 parts) | 94.9 GB | Too big | Get file → |
For 8 GB VRAM, use Q5_K_M (56.11 GB file, ~57.4 GB memory at 8k context). For 16 GB VRAM, use IQ4_XS (93.63 GB file, ~94.9 GB memory at 8k context). Both are split into shards for efficient loading.
What GPU do you need for Qwen3.8 Flash Next 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: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: no quant fits at an 8K context.
Which Qwen3.8 Flash Next Heretic quant should you choose?
Q5_K_M (56.1 GB) is the middle option in this repository. Closer to the original model than 4-bit, for a moderate increase in size. A good pick when you have memory to spare.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3.8 Flash Next Heretic
How much VRAM does Qwen3.8 Flash Next Heretic need?
About 57.4 GB at Q5_K_M with an 8K context window. That is the 56.1 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 Qwen3.8 Flash Next Heretic on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 57.4 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8 Flash Next Heretic GGUF?
It depends on the quant: Q5_K_M is 56.1 GB, IQ4_XS is 93.6 GB.
What context length does Qwen3.8 Flash Next Heretic support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.2 GB at 32K.
Where can I download Qwen3.8 Flash Next Heretic GGUF?
The files are on Hugging Face in the groxaxo/Qwen3.8-Flash-Next-Heretic-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 Qwen3.8 Flash Next 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 this model on 8 GB VRAM?
Yes, using the Q5_K_M quantization. It requires about 57.4 GB memory at 8k context, so you will need to rely on system RAM via mmap for expert weights.
Does this model support vision?
Yes, it includes the F16 mmproj for image input.
What is the native context length?
The model supports 262,144 tokens natively.