ZwZ2-4B needs about 4.3 GB of memory at Q4_K_M (a 2.7 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 3 quants, from 2.7 GB (Q4_K_M) to 8.4 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.
ZwZ2-4B: A 4.21B Qwen3.5-4B model with vision and reasoning
ZwZ2-4B is a 4.21B parameter model based on Qwen3.5-4B. It learns to connect visual evidence across regions through comparison, aggregation, matching, and multi-step reasoning. The model improves overall accuracy from 49.41 to 54.79 on 25 benchmarks across six domains.
Good at
- Vision‑aware reasoning
- Multi‑step multimodal tasks
- Tool‑enabled inference
Limits
- No zooming tools required
- Large context of 262144 tokens
- Apache‑2.0 license
Best for
- Multimodal agents
- Long‑context visual reasoning
- Tool‑augmented workflows
Who it’s for. Developers building multimodal applications that need reasoning over full images without additional zooming steps. The model is suitable for research and production use under an Apache‑2.0 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 2.7 GB | 4.3 GB | Fits | Get file → | |
| Q8_0 | 4.5 GB | 6.1 GB | Fits | Get file → | |
| F16 | 8.4 GB | 10.0 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (2.71 GB, ~4.3 GB memory). For 12 GB VRAM, Q8_0 (4.48 GB, ~6.1 GB memory). For 16 GB VRAM or more, F16 (8.42 GB, ~10 GB memory). Memory estimates are approximate.
What GPU do you need for ZwZ2-4B?
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: the best quant that fits is Q4_K_M (2.7 GB file, about 4.3 GB in use).
- 8 GB VRAM: the best quant that fits is Q8_0 (4.5 GB file, about 6.1 GB in use).
- 12 GB VRAM: the best quant that fits is F16 (8.4 GB file, about 10.0 GB in use).
Which ZwZ2-4B quant should you choose?
Q4_K_M is the usual starting point for ZwZ2-4B: it balances file size and quality well. Here it is a 2.7 GB file.
The Q8_0 file is 4.5 GB, about 1.7 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.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about ZwZ2-4B
How much VRAM does ZwZ2-4B need?
About 4.3 GB at Q4_K_M with an 8K context window. That is the 2.7 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 ZwZ2-4B on an 8 GB GPU?
Yes, with Q8_0: the 4.5 GB file needs about 6.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of ZwZ2-4B GGUF?
It depends on the quant: Q4_K_M is 2.7 GB, Q8_0 is 4.5 GB, F16 is 8.4 GB.
What context length does ZwZ2-4B 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 ZwZ2-4B GGUF?
The files are on Hugging Face in the inclusionAI/ZwZ2-4B-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 ZwZ2-4B 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 ZwZ2-4B on 8 GB VRAM?
Yes, using the Q4_K_M quantized model which requires about 4.3 GB of VRAM at 8 k context.
Does ZwZ2-4B support vision input?
Yes, it includes a vision encoder and projector (mmproj) and can process full images.