Qwen3.8 Flash Next Splash needs about 61.4 GB of memory at its main file (a 54.4 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 8 quants, from 28.8 GB to 92.6 GB (Q5_K_M).
Qwen3.8 Flash Next Splash – 4‑bit and 5‑bit GGUF builds for Apple‑silicon Macs
Qwen3.8 Flash Next Splash offers two GGUF builds, a 4‑bit and a 5‑bit version, each with a choice of PLE table. Both builds fit a 128 GB Mac and include vision support by default. The builds are optimized for the Splash 1.4.0 environment.
Good at
- Efficient 4‑bit inference
- 5‑bit high‑accuracy inference
- Apple‑silicon compatibility
Limits
- Requires 128 GB RAM to run full model
- No explicit context length limit provided
- License type is other
Best for
- Local inference on high‑end Macs
- Vision‑enabled agent workloads
- Testing new Qwen architectures
Who it’s for. Developers and researchers who need a local, high‑capacity Qwen model on Apple‑silicon hardware. The model is suitable for experimentation with the new Qwen architecture and for building vision‑enabled agents.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 28.8 GB | 32.8 GB | Too big | Get file → | |
| — | 28.8 GB | 32.8 GB | Too big | Get file → | |
| — | 54.4 GB | 61.4 GB | Too big | Get file → | |
| — | 54.4 GB | 61.4 GB | Too big | Get file → | |
| — | 82.5 GB (4 parts) | 92.9 GB | Too big | Get file → | |
| Q4_K_XL | 82.5 GB (4 parts) | 92.9 GB | Too big | Get file → | |
| — | 92.6 GB (7 parts) | 104.2 GB | Too big | Get file → | |
| Q5_K_M | 92.6 GB (7 parts) | 104.2 GB | Too big | Get file → |
For a 128 GB Mac, the Q4_K_XL build uses 82.53 GB of memory and the Q5_K_M build uses 92.61 GB. The Q4_K_XL build is smaller and faster, while Q5_K_M offers lower error at the cost of ~10 GB more memory. Choose based on available RAM and accuracy needs.
What GPU do you need for Qwen3.8 Flash Next Splash?
These figures add the file size, an fp16 KV cache at an 8K context window 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: the best quant that fits is the main file (28.8 GB file, about 32.8 GB in use).
Which Qwen3.8 Flash Next Splash quant should you choose?
If memory is tight, the smallest file is the first one in the table at 28.8 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3.8 Flash Next Splash
How much VRAM does Qwen3.8 Flash Next Splash need?
About 61.4 GB at its main file with an 8K context window. That is the 54.4 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 Splash on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 32.8 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8 Flash Next Splash GGUF?
It depends on the quant: file is 28.8 GB, file is 28.8 GB, file is 54.4 GB, file is 54.4 GB, file is 82.5 GB, Q4_K_XL is 82.5 GB and 2 more.
Where can I download Qwen3.8 Flash Next Splash GGUF?
The files are on Hugging Face in the Vardaan97/Qwen3.8-Flash-Next-Splash-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 Splash 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 the 4‑bit build on a machine with 8 GB VRAM?
No. The 4‑bit build requires 82.53 GB of RAM, far exceeding 8 GB VRAM.
Does the model support vision tasks?
Yes, vision is enabled by default in the Splash builds.
What is the context length of this model?
The context length is not specified in the provided facts.