gemma 4 31B it qat heretic japanese imatrix needs about 11.3 GB of memory at IQ2_S (a 9.6 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 6 quants, from 0.0 GB to 16.4 GB (IQ4_XS).
Gemma 4 31B it QAT Japanese imatrix model
Gemma 4 31B it QAT is a 30.7B parameter model optimized for Japanese writing. It is a heavily quantized version of the base Gemma 4 31B it model, released under Apache 2.0. The model is intended for research and safety studies only.
Good at
- Japanese text generation
- Research and alignment studies
- Low‑memory inference
Limits
- No vision support
- Reduced safety alignment
- Requires careful use
Best for
- Japanese language research
- Safety and alignment testing
- Low‑memory deployment
Who it’s for. Researchers and developers who need a Japanese‑optimized Gemma 4 31B model for experimentation. The model is not suitable for public or end‑user services due to its reduced safety alignment.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 0.5 GB | Fits | Get file → | |
| IQ2_XS | 9.0 GB | 10.6 GB | Too big | Get file → | |
| IQ2_S | 9.6 GB | 11.3 GB | Too big | Get file → | |
| Q2_K_XL | 11.8 GB | 13.7 GB | Too big | Get file → | |
| Q3_K_XL | 15.4 GB | 17.7 GB | Too big | Get file → | |
| IQ4_XS | 16.4 GB | 18.8 GB | Too big | Get file → |
For 8 GB VRAM use IQ2_XS (file size 9.02 GB, memory 10.6 GB). For 12 GB VRAM use IQ2_S (file size 9.63 GB, memory 11.3 GB). For 16 GB VRAM use Q2_K_XL (file size 11.77 GB, memory 13.7 GB).
What GPU do you need for gemma 4 31B it qat heretic japanese imatrix?
These figures add the file size, an fp16 KV cache at an 8K context window 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 the main file (0.0 GB file, about 0.5 GB in use).
- 6 GB VRAM: the best quant that fits is the main file (0.0 GB file, about 0.5 GB in use).
- 8 GB VRAM: the best quant that fits is the main file (0.0 GB file, about 0.5 GB in use).
- 12 GB VRAM: the best quant that fits is IQ2_S (9.6 GB file, about 11.3 GB in use).
- 16 GB VRAM: the best quant that fits is Q2_K_XL (11.8 GB file, about 13.7 GB in use).
- 24 GB VRAM: the best quant that fits is IQ4_XS (16.4 GB file, about 18.8 GB in use).
Which gemma 4 31B it qat heretic japanese imatrix quant should you choose?
IQ2_S (9.6 GB) is the middle option in this repository. 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.
If memory is tight, the smallest file is the first one in the table at 0.0 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 gemma 4 31B it qat heretic japanese imatrix
How much VRAM does gemma 4 31B it qat heretic japanese imatrix need?
About 11.3 GB at IQ2_S with an 8K context window. That is the 9.6 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 gemma 4 31B it qat heretic japanese imatrix on an 8 GB GPU?
Yes, with the smaller files: the 0.0 GB file needs about 0.5 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of gemma 4 31B it qat heretic japanese imatrix GGUF?
It depends on the quant: file is 0.0 GB, IQ2_XS is 9.0 GB, IQ2_S is 9.6 GB, Q2_K_XL is 11.8 GB, Q3_K_XL is 15.4 GB, IQ4_XS is 16.4 GB.
Where can I download gemma 4 31B it qat heretic japanese imatrix GGUF?
The files are on Hugging Face in the OS-Software/gemma-4-31B-it-qat-heretic-japanese-imatrix-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 gemma 4 31B it qat heretic japanese imatrix 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 IQ2_XS quant, which requires about 10.6 GB memory at 8k context.
Is this model safe for production use?
No. The model has reduced safety alignment and is intended only for research and experimentation.
What license applies to this model?
Apache 2.0, as stated in the model card.