GE

Gemma-3-12B-It

GemmaGGUF by unslothUpdated 2025-08-14
Download on Hugging Face
Parameters11.8B
Context128K
Layers48
Architecturegemma3
Smallest quant3.1 GB
Quick answer

Gemma-3-12B-It needs about 11.0 GB of memory at Q4_K_M (a 7.3 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 26 quants, from 3.1 GB (IQ1_S) to 23.5 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 12.9 GB.

Gemma‑3‑12B‑It: 12‑billion‑parameter multimodal instruction model

Gemma‑3‑12B‑It is a 12‑billion‑parameter instruction‑tuned model from the Gemma family. It supports text and image input, with a 128K token context window. The model is available in multiple GGUF quantisations for efficient deployment.

Good at

  • Multimodal text and image tasks
  • Large 128K context window
  • Instruction‑tuned responses

Limits

  • Requires 8‑K context for memory estimates
  • No vision support beyond 896x896 images
  • License: gemma

Best for

  • Chatbots needing long context
  • Image‑to‑text summarisation
  • Multilingual Q&A

Who it’s for. Developers building lightweight multimodal applications on laptops or modest cloud instances. The model is suitable for those who need a balance between performance and resource usage.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S3.1 GB 6.8 GB
Fits Get file →
IQ1_M3.2 GB 7.0 GB
Fits Get file →
IQ2_XXS3.6 GB 7.3 GB
Tight Get file →
IQ2_M4.4 GB 8.1 GB
Too big Get file →
Q2_K4.8 GB 8.5 GB
Too big Get file →
Q2_K_L4.8 GB 8.5 GB
Too big Get file →
IQ3_XXS4.8 GB 8.6 GB
Too big Get file →
Q2_K_XL4.9 GB 8.6 GB
Too big Get file →
Q3_K_S5.5 GB 9.2 GB
Too big Get file →
Q3_K_M6.0 GB 9.7 GB
Too big Get file →
Q3_K_XL6.1 GB 9.9 GB
Too big Get file →
IQ4_XS6.6 GB 10.3 GB
Too big Get file →
IQ4_NL6.9 GB 10.6 GB
Too big Get file →
Q4_06.9 GB 10.6 GB
Too big Get file →
Q4_K_S6.9 GB 10.7 GB
Too big Get file →
Q4_K_M7.3 GB 11.0 GB
Too big Get file →
Q4_K_XL7.4 GB 11.2 GB
Too big Get file →
Q4_17.6 GB 11.3 GB
Too big Get file →
Q5_K_S8.2 GB 12.0 GB
Too big Get file →
Q5_K_M8.4 GB 12.2 GB
Too big Get file →
Q5_K_XL8.5 GB 12.2 GB
Too big Get file →
Q6_K9.7 GB 13.4 GB
Too big Get file →
Q6_K_XL10.6 GB 14.3 GB
Too big Get file →
Q8_012.5 GB 16.2 GB
Too big Get file →
Q8_K_XL14.4 GB 18.1 GB
Too big Get file →
BF1623.5 GB 27.3 GB
Too big Get file →

For 8‑GB VRAM, use IQ1_S (3.06 GB, ~6.8 GB RAM). With 12‑GB VRAM, IQ4_K_M (7.3 GB, ~11 GB RAM) is a good balance. For 16‑GB VRAM, Q5_K_XL (8.46 GB, ~12.2 GB RAM) offers higher quality.

What GPU do you need for Gemma-3-12B-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: the best quant that fits is IQ2_XXS (3.6 GB file, about 7.3 GB in use).
  • 12 GB VRAM: the best quant that fits is Q5_K_S (8.2 GB file, about 12.0 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K_XL (10.6 GB file, about 14.3 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_K_XL (14.4 GB file, about 18.1 GB in use).
  • 32 GB VRAM: the best quant that fits is BF16 (23.5 GB file, about 27.3 GB in use).

Which Gemma-3-12B-It quant should you choose?

Q4_K_M is the usual starting point for Gemma-3-12B-It: it balances file size and quality well. Here it is a 7.3 GB file.

The Q8_0 file is 12.5 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 IQ1_S at 3.1 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

Architecturegemma3
Parameters11.8B
Layers48
Trained context131,072 tokens
KV cache (fp16)384 KB per token

Source

Quantized byunsloth
FamilyGemma
Licensegemma
Downloads (HF)64,115
Last updated2025-08-14

Frequently asked questions about Gemma-3-12B-It

How much VRAM does Gemma-3-12B-It need?

About 11.0 GB at Q4_K_M with an 8K context window. That is the 7.3 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 Gemma-3-12B-It on an 8 GB GPU?

Yes, with IQ2_XXS: the 3.6 GB file needs about 7.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Gemma-3-12B-It GGUF?

It depends on the quant: IQ1_S is 3.1 GB, IQ1_M is 3.2 GB, IQ2_XXS is 3.6 GB, IQ2_M is 4.4 GB, Q2_K is 4.8 GB, Q2_K_L is 4.8 GB and 20 more.

What context length does Gemma-3-12B-It support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 3.2 GB at 8K and 12.9 GB at 32K.

Where can I download Gemma-3-12B-It GGUF?

The files are on Hugging Face in the unsloth/gemma-3-12b-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-3-12B-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 Gemma‑3‑12B‑It on 8 GB VRAM?

Yes, using the IQ1_S or IQ1_M quantisations which require about 6.8–7 GB of VRAM at 8‑K context.

Does the model support image input?

Yes, it accepts images resized to 896×896, encoded to 256 tokens each.

What context length does the model support?

The model was trained on 131,072 tokens and supports up to 128K context in inference.

Other GGUFs of gemma-3-12b-it

Similar models

Browse all →

More from unsloth

All by unsloth →

More Gemma models

All Gemma →

Other medium models

All medium models →