Directory › Models › Llama 3.2 3B
Llama 3.2 3B
Llama 3.2 3B is a small, fast Meta model suitable for on-device chat and lightweight tasks. It is a self-hostable AI tool that can be used locally without relying on external data sources. The model is available under the Llama-3.2-Community license.
Who it's for: This model is recommended for users who require a lightweight, on-device chat solution.
Run it locally
ollama run llama3.2:3b
Strengths
- Small and fast
- Suitable for on-device chat and lightweight tasks
- Available under the community-driven Llama-3.2-Community license
Trade-offs
- Limited data sources
- May not be suitable for complex or resource-intensive tasks
Data locality
Runs fully offline — no data leaves your machine.
Measured benchmarks · real hardware
| Hardware | Quant | Speed | VRAM | Notes |
|---|---|---|---|---|
| Apple M4 Max · 36 GB | Q4_K_M | 130.5 tok/s | 5.6 GB | ollama llama3.2:3b |