Gemma 4 A4B 98e v6 Coder It needs about 19.2 GB of memory at Q4_K_M (a 10.6 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 30 quants, from 7.8 GB (IQ2_XS) to 39.8 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 32.2 GB.
Gemma 4 A4B 98e v6 Coder It – 19.9B parameter LLM with code and vision abilities
Gemma 4 A4B 98e v6 Coder It is a 19.9B parameter model that supports coding and vision tasks. It is a mixture of experts with 4 active parameters per expert. The model uses a 262144 token context window.
Good at
- Coding tasks
- Vision inference
- Multi‑turn agentic chat
Limits
- Large VRAM requirement for high‑quality quants
- No native vision support in all quants
- Requires corrected chat template for agentic loops
Best for
- Code generation with tool‑calling
- Image captioning and analysis
- Complex multi‑turn dialogue
Who it’s for. Developers and researchers needing a large, code‑focused model with vision support. The model is suitable for applications that can allocate sufficient GPU memory and use the corrected chat template.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ2_XS | 7.8 GB | 16.3 GB | Too big | Get file → | |
| IQ2_S | 7.8 GB | 16.4 GB | Too big | Get file → | |
| IQ2_M | 8.2 GB | 16.8 GB | Too big | Get file → | |
| IQ3_XXS | 8.9 GB | 17.5 GB | Too big | Get file → | |
| IQ3_XS | 9.2 GB | 17.8 GB | Too big | Get file → | |
| IQ3_XS | 9.3 GB | 17.8 GB | Too big | Get file → | |
| Q3_K_S | 9.7 GB | 18.2 GB | Too big | Get file → | |
| IQ3_XS | 9.7 GB | 18.3 GB | Too big | Get file → | |
| IQ3_M | 9.8 GB | 18.4 GB | Too big | Get file → | |
| — | 10.3 GB | 18.8 GB | Too big | Get file → | |
| Q3_K_M | 10.5 GB | 19.1 GB | Too big | Get file → | |
| Q4_K_M | 10.6 GB | 19.2 GB | Too big | Get file → | |
| Q3_K_XL | 10.7 GB | 19.2 GB | Too big | Get file → | |
| Q3_K_L | 10.9 GB | 19.5 GB | Too big | Get file → | |
| IQ4_XS | 11.0 GB | 19.6 GB | Too big | Get file → | |
| Q4_0 | 11.4 GB | 20.0 GB | Too big | Get file → | |
| IQ4_NL | 11.4 GB | 20.0 GB | Too big | Get file → | |
| Q4_K_S | 12.2 GB | 20.8 GB | Too big | Get file → | |
| Q4_1 | 12.6 GB | 21.2 GB | Too big | Get file → | |
| Q5_K_M | 13.1 GB | 21.6 GB | Too big | Get file → | |
| Q4_K_M | 13.2 GB | 21.8 GB | Too big | Get file → | |
| Q4_K_L | 13.4 GB | 22.0 GB | Too big | Get file → | |
| Q5_K_S | 14.2 GB | 22.7 GB | Too big | Get file → | |
| Q5_K_M | 15.1 GB | 23.6 GB | Too big | Get file → | |
| Q5_K_L | 15.2 GB | 23.8 GB | Too big | Get file → | |
| Q6_K | 15.5 GB | 24.1 GB | Too big | Get file → | |
| Q6_K | 17.8 GB | 26.4 GB | Too big | Get file → | |
| Q6_K_L | 18.0 GB | 26.5 GB | Too big | Get file → | |
| Q8_0 | 21.2 GB | 29.7 GB | Too big | Get file → | |
| F16 | 39.8 GB | 48.3 GB | Too big | Get file → |
For 8 GB VRAM, use IQ2_XS (file size 7.77 GB, estimated 16.3 GB memory at 8k context). For 12 GB VRAM, Q4_K_S (file size 12.21 GB, estimated 20.8 GB memory). For 24 GB VRAM, Q6_K (file size 15.53 GB, estimated 24.1 GB memory). Estimates are approximate.
What GPU do you need for Gemma 4 A4B 98e v6 Coder It?
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: the best quant that fits is Q5_K_L (15.2 GB file, about 23.8 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (21.2 GB file, about 29.7 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (21.2 GB file, about 29.7 GB in use).
Which Gemma 4 A4B 98e v6 Coder It quant should you choose?
Q4_K_M is the usual starting point for Gemma 4 A4B 98e v6 Coder It: it balances file size and quality well. Here it is a 10.6 GB file.
The Q8_0 file is 21.2 GB, about 2.0 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 IQ2_XS at 7.8 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 Gemma 4 A4B 98e v6 Coder It
How much VRAM does Gemma 4 A4B 98e v6 Coder It need?
About 19.2 GB at Q4_K_M with an 8K context window. That is the 10.6 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 A4B 98e v6 Coder It on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 16.3 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Gemma 4 A4B 98e v6 Coder It GGUF?
It depends on the quant: IQ2_XS is 7.8 GB, IQ2_S is 7.8 GB, IQ2_M is 8.2 GB, IQ3_XXS is 8.9 GB, IQ3_XS is 9.2 GB, IQ3_XS is 9.3 GB and 24 more.
What context length does Gemma 4 A4B 98e v6 Coder It support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 8.1 GB at 8K and 32.2 GB at 32K.
Where can I download Gemma 4 A4B 98e v6 Coder It GGUF?
The files are on Hugging Face in the ManniX-ITA/gemma-4-A4B-98e-v6-coder-it-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 A4B 98e v6 Coder It use?
Hugging Face lists the license as gemma. 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 16.3 GB memory at 8k context, so 8 GB VRAM is insufficient.
Does the model support vision tasks?
Yes, the model has detected vision ability, but some quants may not include vision weights.
What is the context length?
The model was trained with a 262144 token context window.