Google · 12B · 密集（Dense）
Gemma 4 mid-size instruct — multimodal any-to-any 看看你的 GPU 或 Mac 跑不跑得動 Gemma 4 12B IT——最低 6.7 GB，建議 11.2 GB。
2026-04256K context
| 量化 | 位元 | VRAM | 品質 | 狀態 |
|---|---|---|---|---|
| Q2_K | 2 | 4.3 GB | low | — |
| Q3_K_M | 3 | 5.9 GB | moderate | — |
| Q4_K_M | 4 | 6.6 GB | good | — |
| Q5_K_M | 5 | 8.2 GB | good | — |
| Q6_K | 6 | 9.7 GB | excellent | — |
| Q8_0 | 8 | 12.8 GB | excellent | — |
| F16 | 16 | 25.1 GB | lossless | — |
Can I run Gemma 4 12B IT locally?
- Can I run Gemma 4 12B IT locally?
- Gemma 4 12B IT needs about 6.7 GB of memory at a minimum and 11.2 GB recommended. Open this page to grade it against your GPU or Mac, then run it with runai, Ollama or LM Studio.
- How much VRAM does Gemma 4 12B IT need?
- At Q4_K_M, Gemma 4 12B IT uses about 6.6 GB of VRAM. Higher quants need more memory; lower quants fit tighter cards with a quality tradeoff.