spec for testing needs about 1.0 GB of memory at Q4_K_M (a 0.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 9 quants, from 0.2 GB (Q8_0) to 22.0 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 0.7 GB.
Speculative decoding Qwen3-4B model in GGUF format
Spec for testing is a Qwen3-4B derived model using dflash architecture. It supports reasoning and tool usage. The model offers a 40960 token context length and is available in multiple quantizations.
Good at
- Reasoning
- Tool integration
- Large context
Limits
- No vision support
- License is other
- No mixture of experts
Best for
- Chatbot prototypes
- Tool‑enabled agents
- Long‑form reasoning
Who it’s for. Developers building lightweight LLM applications who need a 4B‑parameter model with speculative decoding support. The model is useful for testing llama.cpp integration and for educational purposes.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q8_0 | 0.2 GB | 0.9 GB | Fits | Get file → | |
| Q4_K_M | 0.3 GB | 1.0 GB | Fits | Get file → | |
| Q4_K_M | 0.6 GB | 1.2 GB | Fits | Get file → | |
| — | 1.0 GB | 1.7 GB | Fits | Get file → | |
| Q8_0 | 2.1 GB | 2.7 GB | Fits | Get file → | |
| Q4_K_M | 2.5 GB | 3.2 GB | Fits | Get file → | |
| Q4_K_M | 3.1 GB | 3.8 GB | Fits | Get file → | |
| Q4_K_M | 19.4 GB | 20.1 GB | Too big | Get file → | |
| — | 22.0 GB | 22.7 GB | Too big | Get file → |
For 8 GB VRAM, use Q8_0 (0.24 GB file, 0.9 GB memory). For 4 GB VRAM, use Q4_K_M (0.31 GB file, 1 GB memory). For 2 GB VRAM, use Q4_K_M (0.57 GB file, 1.2 GB memory). Memory estimates are approximate.
What GPU do you need for spec for testing?
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 Q4_K_M (3.1 GB file, about 3.8 GB in use).
- 6 GB VRAM: the best quant that fits is Q4_K_M (3.1 GB file, about 3.8 GB in use).
- 8 GB VRAM: the best quant that fits is Q4_K_M (3.1 GB file, about 3.8 GB in use).
- 12 GB VRAM: the best quant that fits is Q4_K_M (3.1 GB file, about 3.8 GB in use).
- 16 GB VRAM: the best quant that fits is Q4_K_M (3.1 GB file, about 3.8 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (22.0 GB file, about 22.7 GB in use).
Which spec for testing quant should you choose?
Q4_K_M is the usual starting point for spec for testing: it balances file size and quality well. Here it is a 0.3 GB file.
The Q8_0 file is 0.2 GB, about 0.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is Q8_0 at 0.2 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about spec for testing
How much VRAM does spec for testing need?
About 1.0 GB at Q4_K_M 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 spec for testing on an 8 GB GPU?
Yes, with Q4_K_M: the 3.1 GB file needs about 3.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of spec for testing GGUF?
It depends on the quant: Q8_0 is 0.2 GB, Q4_K_M is 0.3 GB, Q4_K_M is 0.6 GB, file is 1.0 GB, Q8_0 is 2.1 GB, Q4_K_M is 2.5 GB and 3 more.
What context length does spec for testing support?
The GGUF header lists a trained context length of 40,960 tokens (40K). The KV cache grows with the context you use: about 0.2 GB at 8K and 0.7 GB at 32K.
Where can I download spec for testing GGUF?
The files are on Hugging Face in the ggml-org/spec-for-testing-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 spec for testing 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?
Yes, the Q8_0 quantization requires about 0.9 GB of VRAM at 8 k context, which fits within 8 GB.
Does this model support vision tasks?
No, the model has no vision capabilities.
What license applies to these GGUF files?
The files retain the licenses of their source models; review each source model card for details.