Nova-v3 needs about 1.1 GB of memory at F16 (a 0.3 GB file) with an 8K context window, so it fits on a 4 GB GPU.
Nova‑v3: 148.8 million‑parameter decoder‑only language model
Nova‑v3 is a 148.8 million‑parameter decoder‑only causal language model. It supports key/value caching and a 2,048‑token context window. The GGUF export uses F16 weights and is available for llama.cpp and Apple Metal.
Good at
- Coding tasks
- Causal language generation
- Key/value caching
Limits
- 2,048‑token context limit
- No vision support
- Apache‑2.0 license
Best for
- Small‑scale code generation
- Mobile inference on Apple devices
- Low‑memory GPU usage
Who it’s for. Developers needing a lightweight, open‑source language model for coding or casual text generation. It is suitable for mobile or embedded applications that can run GGUF files with llama.cpp.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| F16 | 0.3 GB | 0.9 GB | Fits | Get file → |
The only available quant is F16, 0.34 GB in size. It requires approximately 1.1 GB of VRAM at 8,000‑token context. This is suitable for GPUs with 8 GB or more of VRAM. No other quantizations are provided.
What GPU do you need for Nova-v3?
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: the best quant that fits is F16 (0.3 GB file, about 1.1 GB in use).
Which Nova-v3 quant should you choose?
F16 (0.3 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Nova-v3
How much VRAM does Nova-v3 need?
About 1.1 GB at F16 with an 8K context window. That is the 0.3 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 Nova-v3 on an 8 GB GPU?
Yes, with F16: the 0.3 GB file needs about 1.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Nova-v3 GGUF?
It depends on the quant: F16 is 0.3 GB.
What context length does Nova-v3 support?
The GGUF header lists a trained context length of 2,048 tokens (2K). The KV cache grows with the context you use: about 0.3 GB at 8K and 0.1 GB at 2K.
Where can I download Nova-v3 GGUF?
The files are on Hugging Face in the plasmova/nova-v3 repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Nova-v3 use?
Hugging Face lists the license as apache-2.0. 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 Nova‑v3 on 8 GB VRAM?
Yes. The F16 GGUF file uses about 1.1 GB of VRAM at 8,000‑token context, well within an 8 GB GPU.
Does Nova‑v3 support vision tasks?
No, Nova‑v3 is a text‑only decoder‑only model.