← Back to Models
Gemma 3
Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. The 4B, 12B, and 27B models can process both text and image inputs.
gemma-3-4b
4B parameters
Base
Pull this model
Use the following command with the HoML CLI:
homl pull gemma-3:4bResource Requirements
| Quantization | Disk Space | GPU Memory |
|---|---|---|
| BF16 | 8 GB | 8 GB |
gemma-3-4b-it
4B parameters
Instruction-Tuned
Pull this model
Use the following command with the HoML CLI:
homl pull gemma-3:4b-itResource Requirements
| Quantization | Disk Space | GPU Memory |
|---|---|---|
| BF16 | 8 GB | 8 GB |
gemma-3-12b
12B parameters
Base
Pull this model
Use the following command with the HoML CLI:
homl pull gemma-3:12bResource Requirements
| Quantization | Disk Space | GPU Memory |
|---|---|---|
| BF16 | 24 GB | 24 GB |
gemma-3-12b-it
12B parameters
Instruction-Tuned
Pull this model
Use the following command with the HoML CLI:
homl pull gemma-3:12b-itResource Requirements
| Quantization | Disk Space | GPU Memory |
|---|---|---|
| BF16 | 24 GB | 24 GB |
gemma-3-27b
27B parameters
Base
Pull this model
Use the following command with the HoML CLI:
homl pull gemma-3:27bResource Requirements
| Quantization | Disk Space | GPU Memory |
|---|---|---|
| BF16 | 54 GB | 54 GB |
gemma-3-27b-it
27B parameters
Instruction-Tuned
Pull this model
Use the following command with the HoML CLI:
homl pull gemma-3:27b-itResource Requirements
| Quantization | Disk Space | GPU Memory |
|---|---|---|
| BF16 | 54 GB | 54 GB |