WARD 2b needs about 2.5 GB of memory at Q4_K_M (a 1.6 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 12 quants, from 1.1 GB (Q2_K) to 4.8 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 1.6 GB.
WARD 2b – multimodal guard model for prompt injection detection
WARD 2b is a 2.39‑billion parameter multimodal model built on qwen35. It detects prompt injections from HTML text and screenshots, returning JSON with reasoning, attack_goal, injection_location, and label. The model is released under apache‑2.0.
Good at
- HTML prompt injection detection
- Screenshot analysis
- Reasoning output
Limits
- No vision support beyond screenshots
- Requires 8k context for estimates
- License is apache‑2.0
Best for
- Web agent security checks
- Content moderation pipelines
- Automated injection alerts
Who it’s for. Developers building web agents who need to guard against prompt injections. The model is suitable for security teams and compliance auditors working with HTML and visual inputs.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 1.1 GB | 2.0 GB | Fits | Get file → | |
| Q3_K_S | 1.2 GB | 2.1 GB | Fits | Get file → | |
| Q3_K_M | 1.3 GB | 2.2 GB | Fits | Get file → | |
| Q3_K_L | 1.4 GB | 2.3 GB | Fits | Get file → | |
| IQ4_XS | 1.5 GB | 2.4 GB | Fits | Get file → | |
| Q4_K_S | 1.5 GB | 2.4 GB | Fits | Get file → | |
| Q4_K_M | 1.6 GB | 2.5 GB | Fits | Get file → | |
| Q5_K_S | 1.7 GB | 2.6 GB | Fits | Get file → | |
| Q5_K_M | 1.8 GB | 2.7 GB | Fits | Get file → | |
| Q6_K | 2.0 GB | 2.9 GB | Fits | Get file → | |
| Q8_0 | 2.6 GB | 3.5 GB | Fits | Get file → | |
| F16 | 4.8 GB | 5.7 GB | Fits | Get file → |
For 8 GB VRAM, use Q4_K_M (1.56 GB, ~2.5 GB memory). If 4 GB VRAM is available, Q3_K_M (1.32 GB, ~2.2 GB memory) is a good balance. For larger budgets, Q8_0 (2.55 GB, ~3.5 GB memory) offers higher precision.
What GPU do you need for WARD 2b?
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 Q8_0 (2.6 GB file, about 3.5 GB in use).
- 6 GB VRAM: the best quant that fits is F16 (4.8 GB file, about 5.7 GB in use).
Which WARD 2b quant should you choose?
Q4_K_M is the usual starting point for WARD 2b: it balances file size and quality well. Here it is a 1.6 GB file.
The Q8_0 file is 2.6 GB, about 1.6 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 Q2_K at 1.1 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about WARD 2b
How much VRAM does WARD 2b need?
About 2.5 GB at Q4_K_M with an 8K context window. That is the 1.6 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 WARD 2b on an 8 GB GPU?
Yes, with F16: the 4.8 GB file needs about 5.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of WARD 2b GGUF?
It depends on the quant: Q2_K is 1.1 GB, Q3_K_S is 1.2 GB, Q3_K_M is 1.3 GB, Q3_K_L is 1.4 GB, IQ4_XS is 1.5 GB, Q4_K_S is 1.5 GB and 6 more.
What context length does WARD 2b support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.4 GB at 8K and 1.6 GB at 32K.
Where can I download WARD 2b GGUF?
The files are on Hugging Face in the mradermacher/WARD-2b-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 WARD 2b 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 WARD 2b on 8 GB VRAM?
Yes, the Q4_K_M quant uses 1.56 GB and requires about 2.5 GB VRAM at 8k context.
Does WARD 2b support vision?
It processes screenshots but does not have full vision capabilities beyond that.