Qwen3.5 2B Heretic needs about 2.2 GB of memory at Q4_K_M (a 1.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 3 quants, from 1.3 GB (Q4_K_M) to 3.9 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 1.7 GB.
Qwen3.5 2B Heretic – a 1.94B parameter model with reduced refusals and vision support
Qwen3.5 2B Heretic removes refusal behaviour via Arbitrary‑Rank Ablation while retaining full weight and vision encoder. It supports reasoning, tools, and vision tasks. The model is released under apache‑2.0 and can be used with transformers, vLLM or SGLang.
Good at
- Reasoning and multi‑token prediction
- Vision description
- Tool usage
Limits
- No vision support without mmproj
- Small context size limits
- License requires compliance
Best for
- Chatbots needing low refusal rates
- Image captioning with text prompts
- Tool‑enabled workflows
Who it’s for. Developers building applications that require a lightweight Qwen model with reduced safety guardrails. The model is suitable for research and production use where the user accepts the responsibility for its outputs.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 1.3 GB | 2.2 GB | Fits | Get file → | |
| Q8_0 | 2.1 GB | 3.0 GB | Fits | Get file → | |
| BF16 | 3.9 GB | 4.8 GB | Fits | Get file → |
For 8 GB VRAM use Q4_K_M (1.31 GB file, ~2.2 GB memory at 8 k context). For 12 GB VRAM use Q8_0 (2.08 GB file, ~3 GB memory). For 16 GB VRAM use BF16 (3.9 GB file, ~4.8 GB memory). Memory estimates are approximate.
What GPU do you need for Qwen3.5 2B Heretic?
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 (2.1 GB file, about 3.0 GB in use).
- 6 GB VRAM: the best quant that fits is BF16 (3.9 GB file, about 4.8 GB in use).
Which Qwen3.5 2B Heretic quant should you choose?
Q4_K_M is the usual starting point for Qwen3.5 2B Heretic: it balances file size and quality well. Here it is a 1.3 GB file.
The Q8_0 file is 2.1 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.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3.5 2B Heretic
How much VRAM does Qwen3.5 2B Heretic need?
About 2.2 GB at Q4_K_M with an 8K context window. That is the 1.3 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 Qwen3.5 2B Heretic on an 8 GB GPU?
Yes, with BF16: the 3.9 GB file needs about 4.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Qwen3.5 2B Heretic GGUF?
It depends on the quant: Q4_K_M is 1.3 GB, Q8_0 is 2.1 GB, BF16 is 3.9 GB.
What context length does Qwen3.5 2B Heretic 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 1.7 GB at 32K.
Where can I download Qwen3.5 2B Heretic GGUF?
The files are on Hugging Face in the darrellbest/Qwen3.5-2B-Heretic-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 Qwen3.5 2B Heretic 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 Qwen3.5 2B Heretic on 8 GB VRAM?
Yes, the Q4_K_M quantized file is 1.31 GB and requires about 2.2 GB VRAM at 8 k context.
Does the model support image input?
Yes, the mmproj file carries the vision encoder and must be loaded alongside any GGUF file.
What license applies to this model?
The model is released under the apache‑2.0 license.