Models / gemma-3
gemma-3 by unsloth
7 models (270M to 27B) · ↓ 256.2K downloads a month · from $0.10/h on RTX 3060 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
gemma-3 270M
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalunsloth/gemma-3-270m-it | vLLM | 540 MB | 3 GB | $0.10/h RTX 3060 |
gemma-3 270M
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalunsloth/gemma-3-270m | vLLM | 540 MB | 3 GB | $0.10/h RTX 3060 |
gemma-3 1B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalunsloth/gemma-3-1b-it | vLLM | 2.0 GB | 4.7 GB | $0.10/h RTX 3060 |
gemma-3 1B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalunsloth/gemma-3-1b-pt | vLLM | 2.0 GB | 4.7 GB | $0.10/h RTX 3060 |
gemma-3 4B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalunsloth/gemma-3-4b-it | vLLM | 8.6 GB | 13 GB | $0.14/h RTX A4000 |
gemma-3 12B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Onnx (Prince-1)Prince-1/Gemma-3-12b-pt-Onnx | ONNX Runtime GenAI | 26 GB | 34.3 GB | $0.32/h Quadro RTX 8000 |
gemma-3 27B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalunsloth/gemma-3-27b-it | vLLM | 55 GB | 67.8 GB | $0.89/h 2x Nvidia A40 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").