Hy MT2 30B A3B Uncensored Heretic needs about 12.6 GB of memory at its main file (a 11.3 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 4 quants, from 9.7 GB to 19.1 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.
30B multilingual translation model with uncensored behavior
Hy MT2 30B A3B Uncensored Heretic is a 30‑billion‑parameter multilingual translation model. It is a mixture of experts with 3 billion active parameters and supports 262,144‑token context. The model is released under Apache‑2.0 and is intended for research and safety studies.
Good at
- Large‑context multilingual translation
- High‑capacity inference
- Research‑grade uncensored output
Limits
- No vision support
- Requires large VRAM for full precision
- Uncensored content may be unsafe
Best for
- Multilingual translation research
- Safety and alignment testing
- Red‑team evaluation
Who it’s for. Researchers, safety engineers, and developers needing a large‑scale translation model for experimentation. The model is not suited for public deployment due to its uncensored nature.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 9.7 GB | 11.0 GB | Too big | Get file → | |
| — | 11.3 GB | 12.6 GB | Too big | Get file → | |
| — | 14.0 GB | 15.3 GB | Too big | Get file → | |
| — | 19.1 GB | 20.4 GB | Too big | Get file → |
For 8 GB VRAM use the 9.66 GB file (est. 11 GB memory at 8 k context). For 12 GB VRAM use the 11.27 GB file (est. 12.6 GB). For 16 GB VRAM use the 14.04 GB file (est. 15.3 GB). For 24 GB VRAM use the 19.07 GB file (est. 20.4 GB).
What GPU do you need for Hy MT2 30B A3B Uncensored 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: 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: the best quant that fits is the main file (9.7 GB file, about 11.0 GB in use).
- 16 GB VRAM: the best quant that fits is the main file (14.0 GB file, about 15.3 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (19.1 GB file, about 20.4 GB in use).
Which Hy MT2 30B A3B Uncensored Heretic quant should you choose?
If memory is tight, the smallest file is the first one in the table at 9.7 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Hy MT2 30B A3B Uncensored Heretic
How much VRAM does Hy MT2 30B A3B Uncensored Heretic need?
About 12.6 GB at its main file with an 8K context window. That is the 11.3 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 Hy MT2 30B A3B Uncensored Heretic on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 11.0 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Hy MT2 30B A3B Uncensored Heretic GGUF?
It depends on the quant: file is 9.7 GB, file is 11.3 GB, file is 14.0 GB, file is 19.1 GB.
What context length does Hy MT2 30B A3B Uncensored 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 0.8 GB at 8K and 3.2 GB at 32K.
Where can I download Hy MT2 30B A3B Uncensored Heretic GGUF?
The files are on Hugging Face in the alphaZimuth/Hy-MT2-30B-A3B-Uncensored-Heretic-APEX-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 Hy MT2 30B A3B Uncensored 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 this model on 8 GB VRAM?
Use the 9.66 GB quantized file; it requires about 11 GB of VRAM at 8 k context, which is slightly above 8 GB.
Is this model safe for public use?
No. The model is uncensored and may produce harmful content. Use only in controlled research settings.
What license applies to this model?
Apache‑2.0. The model may be redistributed under the same license terms.