Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 needs about 61.8 GB of memory at IQ3_XXS (a 60.5 GB file) with an 8K context window, which is more than a 48 GB GPU holds. Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.
Sakura Qwen3.8 Flash-Next Compact ISTA IQ3 XXS 512E CP2 – 177B parameters, 512 experts
Sakura Qwen3.8 Flash-Next Compact ISTA IQ3 XXS 512E CP2 is a 177‑billion‑parameter Qwen model quantized to IQ3_XXS. It retains all 512 routed experts per layer across 48 layers. The GGUF file is split into two shards and totals 60.48 GB.
Good at
- Reasoning
- Tools usage
- Large‑expert routing
Limits
- Requires ~61.8 GB VRAM at 8 k context
- Model size 60.48 GB – large download
- Experimental release – not tested for inference
Best for
- High‑capacity inference on large‑GPU servers
- Research on sparse MoE models
- Benchmarking non‑uniform quantization
Who it’s for. Researchers and developers needing a large‑parameter Qwen model with sparse expert routing and non‑uniform IQ3_XXS quantization. The model is suitable for environments with ample GPU memory and a need for experimental structural releases.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ3_XXS | 60.5 GB (2 parts) | 61.8 GB | Too big | Get file → |
For a 8 k context window, the IQ3_XXS quantization requires about 61.8 GB of VRAM. If you have a 16 GB GPU, this model will not fit. Consider using a smaller quantization or a different model if memory is limited.
What GPU do you need for Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2?
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 Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 quant should you choose?
IQ3_XXS (60.5 GB) is the middle option in this repository. 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
Frequently asked questions about Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2
How much VRAM does Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 need?
About 61.8 GB at IQ3_XXS with an 8K context window. That is the 60.5 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 Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 61.8 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 GGUF?
It depends on the quant: IQ3_XXS is 60.5 GB.
What context length does Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 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 Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 GGUF?
The files are on Hugging Face in the webmp3/Sakura-Qwen3.8-Flash-Next-Compact-ISTA-IQ3_XXS-512E-CP2-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 Sakura Qwen3.8 Flash Next Compact ISTA IQ3 XXS 512E CP2 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?
No. The model requires approximately 61.8 GB of VRAM at an 8 k context length.
Does this model support vision tasks?
The release notes do not confirm vision support; it is not guaranteed to work with multimodal inputs.
Is this a mixture‑of‑experts model?
Yes, it uses 512 routed experts per layer with 10 active experts per token.