StartLux Decision v1 9B BF16 needs about 19.5 GB of memory at BF16 (a 17.9 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 4.3 GB.
StartLux Decision v1 9B BF16 – a decision‑making LLM with vision support
StartLux Decision v1 9B BF16 is a 8.95 billion‑parameter model built on the qwen35 architecture. It answers typed questions about a state, which can be text, JSON or images, up to 262,144 tokens. The model returns a probability for each option and uses a TypeSafe API format.
Good at
- Fast inference with one forward pass per request
- Supports image evidence in prompts
- Handles up to 262,144‑token contexts
Limits
- Requires BF16 precision; no other quantizations listed
- CC BY‑NC 4.0 license limits commercial use
- No explicit vision‑only mode; vision tower included but not separate
Best for
- Rapid decision‑making in QA systems
- Batching multiple questions for high throughput
- Integrating image‑based evidence in prompts
Who it’s for. Researchers and developers needing a decision‑oriented LLM that can process large contexts and images. Suitable for non‑commercial projects under a CC BY‑NC 4.0 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| BF16 | 17.9 GB | 19.5 GB | Too big | Get file → |
The only available quant is BF16, 17.92 GB file size and 19.5 GB estimated VRAM at 8 k context. For 8 GB VRAM, the model cannot run in full precision; consider a lower‑precision quant if available. For 16 GB VRAM, BF16 is feasible with the estimated memory usage.
What GPU do you need for StartLux Decision v1 9B BF16?
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 BF16 (17.9 GB file, about 19.5 GB in use).
Which StartLux Decision v1 9B BF16 quant should you choose?
BF16 (17.9 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 StartLux Decision v1 9B BF16
How much VRAM does StartLux Decision v1 9B BF16 need?
About 19.5 GB at BF16 with an 8K context window. That is the 17.9 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 StartLux Decision v1 9B BF16 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 19.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of StartLux Decision v1 9B BF16 GGUF?
It depends on the quant: BF16 is 17.9 GB.
What context length does StartLux Decision v1 9B BF16 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.
Where can I download StartLux Decision v1 9B BF16 GGUF?
The files are on Hugging Face in the startlux-models/StartLux-Decision-v1-9B-BF16-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 StartLux Decision v1 9B BF16 use?
Hugging Face lists the license as cc-by-nc-4.0. 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 this model on 8 GB VRAM?
The BF16 quant requires 19.5 GB VRAM at 8 k context, so 8 GB is insufficient.
Does the model support vision input?
Yes, the weights include a vision tower and the inference package accepts images in the request.
Is commercial use allowed?
The model is CC BY‑NC 4.0; commercial use requires a separate license from StartLux Labs.