Gemma 3 12B — Specifications
| Developer | |
|---|---|
| Type | LLM (multimodal) |
| Modality | Text, Image → Text |
| Parameters | 12B |
| Context window | 128K |
| Max output | — |
| License | Gemma (open) |
| Open weights | ✅ Yes |
| Released | 2025 |
| Input price | $0.05 /1M |
| Output price | $0.15 /1M |
| API providers | Google AI Studio, OpenRouter, Ollama |
🖥️ Run it locally
| VRAM (4-bit) | ~8 GB |
|---|---|
| Minimum GPU | RTX 4070 12GB |
A mid-size multimodal Gemma 3 (text + image, 128K, 140+ languages) that runs 4-bit on a 12GB GPU — a great balance of capability and footprint.