Gemma 4 MoonGem 31B needs about 52.2 GB of memory at Q4_K_M (a 19.5 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 26 quants, from 0.0 GB to 61.4 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 128.8 GB.
Gemma 4 MoonGem 31B – 31 B‑parameter multimodal LLM
Gemma 4 MoonGem 31B is a 31 B‑parameter model that supports text, image and audio input. It is quantized for llama.cpp and offers a range of precision options. The model is released under Apache‑2.0 and is available on Hugging Face.
Good at
- Multimodal text, image and audio processing
- Large context support up to 262 144 tokens
- Role‑play and conversational tasks
Limits
- Requires 8 k context for memory estimates
- No speculative decoding
- No vision support beyond image input
Best for
- Chatbots needing multimodal input
- Content creation with image prompts
- Audio‑to‑text generation
Who it’s for. Developers building multimodal chat applications or content tools that need a large‑parameter model with flexible quantization. The model is suitable for users with sufficient GPU memory to run the chosen quantization.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 32.7 GB | Too big | Get file → | |
| IQ2_XXS | 9.2 GB | 41.9 GB | Too big | Get file → | |
| IQ2_XS | 9.6 GB | 42.3 GB | Too big | Get file → | |
| IQ2_S | 10.2 GB | 42.9 GB | Too big | Get file → | |
| IQ2_M | 11.3 GB | 44.1 GB | Too big | Get file → | |
| Q2_K | 11.9 GB | 44.6 GB | Too big | Get file → | |
| IQ3_XXS | 13.3 GB | 46.0 GB | Too big | Get file → | |
| Q3_K_S | 14.0 GB | 46.7 GB | Too big | Get file → | |
| IQ3_XS | 14.0 GB | 46.7 GB | Too big | Get file → | |
| Q3_K_M | 14.7 GB | 47.4 GB | Too big | Get file → | |
| Q3_K_L | 15.6 GB | 48.3 GB | Too big | Get file → | |
| IQ3_M | 16.4 GB | 49.1 GB | Too big | Get file → | |
| IQ4_XS | 17.2 GB | 49.9 GB | Too big | Get file → | |
| Q4_0 | 18.1 GB | 50.8 GB | Too big | Get file → | |
| Q4_K_S | 18.3 GB | 51.0 GB | Too big | Get file → | |
| IQ4_NL | 19.4 GB | 52.1 GB | Too big | Get file → | |
| Q4_K_M | 19.5 GB | 52.2 GB | Too big | Get file → | |
| Q4_1 | 19.8 GB | 52.5 GB | Too big | Get file → | |
| Q4_K_L | 20.9 GB | 53.6 GB | Too big | Get file → | |
| Q5_K_S | 22.0 GB | 54.7 GB | Too big | Get file → | |
| Q5_K_M | 23.5 GB | 56.2 GB | Too big | Get file → | |
| Q6_K_S | 25.8 GB | 58.5 GB | Too big | Get file → | |
| Q6_K | 27.0 GB | 59.7 GB | Too big | Get file → | |
| Q6_K_L | 28.3 GB | 61.0 GB | Too big | Get file → | |
| Q8_0 | 32.6 GB | 65.3 GB | Too big | Get file → | |
| BF16 | 61.4 GB (2 parts) | 94.1 GB | Too big | Get file → |
For 8 GB VRAM, use IQ2_XXS (9.16 GB) or IQ2_XS (9.61 GB). For 16 GB VRAM, Q4_K_M (19.52 GB) or Q4_K_S (18.31 GB) provide a good balance of size and performance. Higher‑precision quants such as Q6_K (26.96 GB) require more memory but offer better quality.
What GPU do you need for Gemma 4 MoonGem 31B?
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: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: the best quant that fits is Q3_K_M (14.7 GB file, about 47.4 GB in use).
Which Gemma 4 MoonGem 31B quant should you choose?
Q4_K_M is the usual starting point for Gemma 4 MoonGem 31B: it balances file size and quality well. Here it is a 19.5 GB file.
The Q8_0 file is 32.6 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.
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 MoonGem 31B
How much VRAM does Gemma 4 MoonGem 31B need?
About 52.2 GB at Q4_K_M with an 8K context window. That is the 19.5 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 MoonGem 31B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 32.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Gemma 4 MoonGem 31B GGUF?
It depends on the quant: file is 0.0 GB, IQ2_XXS is 9.2 GB, IQ2_XS is 9.6 GB, IQ2_S is 10.2 GB, IQ2_M is 11.3 GB, Q2_K is 11.9 GB and 20 more.
What context length does Gemma 4 MoonGem 31B support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 32.2 GB at 8K and 128.8 GB at 32K.
Where can I download Gemma 4 MoonGem 31B GGUF?
The files are on Hugging Face in the bartowski/Ateron_Gemma-4-MoonGem-31B-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 MoonGem 31B 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 Gemma 4 MoonGem 31B on 8 GB VRAM?
Use the IQ2_XXS or IQ2_XS quantization; both fit within 8 GB VRAM with approximate memory usage of 41–42 GB at 8 k context.
Does the model support image input?
Yes, it supports image input via the mmproj projector files included in the repository.
What license applies to this model?
The model is released under the Apache‑2.0 license.