Swift 1.5 Qwen3.8 27B NVFP4 needs about 22.6 GB of memory at its main file (a 19.9 GB file) with an 8K context window, so it fits on a 24 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.
Swift 1.5 Qwen3.8 27B NVFP4 – a reasoning‑efficient Qwen model
Swift 1.5 Qwen3.8 27B NVFP4 is a 27.3‑billion‑parameter model derived from Qwen3.8‑27B. It uses NVIDIA NVFP4/FP8 mixed precision and is optimized for reasoning and tool use. The model is available as a llama.cpp GGUF file without requantization.
Good at
- Reasoning tasks
- Tool usage
- Vision support
Limits
- Requires ModelOpt‑compatible runtime
- No native NVFP4 support on non‑Blackwell GPUs
- Large VRAM requirement
Best for
- Agentic applications
- Complex coding assistance
- Vision‑enabled workflows
Who it’s for. Developers and researchers needing a high‑capacity reasoning model with efficient token usage. The model is suitable for environments that support NVIDIA mixed‑precision runtimes.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 19.9 GB | 22.6 GB | Too big | Get file → |
The only available quant is "unknown" with a file size of 19.92 GB. It requires approximately 22.6 GB of VRAM at an 8 k context length. For lower memory budgets, a different quantization would be needed, but none are provided in this release.
What GPU do you need for Swift 1.5 Qwen3.8 27B NVFP4?
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 the main file (19.9 GB file, about 22.6 GB in use).
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Swift 1.5 Qwen3.8 27B NVFP4
How much VRAM does Swift 1.5 Qwen3.8 27B NVFP4 need?
About 22.6 GB at its main file with an 8K context window. That is the 19.9 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 Swift 1.5 Qwen3.8 27B NVFP4 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 22.6 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Swift 1.5 Qwen3.8 27B NVFP4 GGUF?
It depends on the quant: file is 19.9 GB.
What context length does Swift 1.5 Qwen3.8 27B NVFP4 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.2 GB at 8K and 8.7 GB at 32K.
Where can I download Swift 1.5 Qwen3.8 27B NVFP4 GGUF?
The files are on Hugging Face in the HuggingJoost/Swift-1.5-Qwen3.8-27B-NVFP4-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 Swift 1.5 Qwen3.8 27B NVFP4 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 Swift 1.5 Qwen3.8 27B on 8 GB VRAM?
No. The model requires about 22.6 GB of VRAM at 8 k context, which exceeds an 8 GB budget.
Does this model support vision tasks?
Yes, vision is listed among its detected abilities.
What runtime is needed for NVFP4 execution?
A ModelOpt‑compatible runtime such as vLLM or TensorRT‑LLM is required; native NVFP4 execution needs NVIDIA Blackwell support.