Google DeepMind · gemma3n
Can I run gemma3n?
Gemma 3n models are designed for efficient execution on everyday devices such as laptops, tablets or phones.
6.9B parameters32K contexttextGemma Terms of Use6 builds
gemma3n system requirements
Memory needed at 4K context, counting model weights, KV cache, compute buffers and runtime overhead. Download sizes come from the official Ollama registry.
| Build | Quantization | Download | Memory needed |
|---|---|---|---|
| gemma3n:e2b-it-q4_K_M | Q4_K_M | 5.2 GB | ~7.9 GB(est.) |
| gemma3n:e2b-it-q8_0 | Q8_0 | 6.2 GB | ~9.2 GB(est.) |
| gemma3n:e4b-it-q4_K_M | Q4_K_M | 7.0 GB | ~10.4 GB(est.) |
| gemma3n:e2b-it-fp16 | F16 | 8.3 GB | ~12.3 GB(est.) |
| gemma3n:e4b-it-q8_0 | Q8_0 | 8.9 GB | ~13.1 GB(est.) |
| gemma3n:e4b-it-fp16 | F16 | 12.8 GB | ~18.7 GB(est.) |
Figures marked (est.) approximate the KV cache because this model's architecture details are not published. They are indicative rather than exact.
How to run gemma3n locally
- 1.Install Ollama for Windows, macOS or Linux.
- 2.Run this in a terminal:
ollama run gemma3n:e2b-it-q4_K_M - 3.The weights download on first run, then an interactive prompt opens.
Frequently asked questions about gemma3n
- How much VRAM do I need to run gemma3n?
- At 4K context, the smallest published build of gemma3n (Q4_K_M) needs roughly 7.9 GB of GPU memory once weights, KV cache, compute buffers and runtime overhead are counted. Bigger quantizations and longer context windows need more. It can also run partly or entirely on the CPU using system RAM, more slowly.
- Can I run gemma3n without a dedicated GPU?
- Yes, but slowly. Without a GPU the model runs on the CPU using system RAM, which usually means a few tokens per second rather than dozens. You would need at least 7.9 GB of free RAM for the smallest build.
- How big is the gemma3n download?
- The smallest published build is 5.2 GB. There are 6 builds in total, the largest being 12.8 GB. Leave some extra free disk space beyond the download itself.
- How do I run gemma3n locally?
- Install Ollama, then run "ollama run gemma3n:e2b-it-q4_K_M" in a terminal. The weights download on first use and an interactive session opens.
Will gemma3n run on your PC?
Scan your hardware once and get a verdict for this model and every build — free, no account.
Model data from the official Ollama registry · verified 13/09/2026