Overview
Llama 3.2 on Ollama brings Meta's efficient models to your local machine. Available in 1B, 3B, and 11B sizes, these models are optimized for fast inference on consumer hardware including MacBooks and laptops.
Sizes
1B / 3B / 11B
Context
128K tokens
VRAM
1-8GB
License
Open Weights
✅Strengths
- ✓Runs on MacBook Air (1B/3B models)
- ✓Fast inference for real-time apps
- ✓128K context in efficient package
- ✓Vision support in 11B and 90B variants
- ✓Free, private, no API costs
⚠️Weaknesses
- ✗Smaller models have limited reasoning
- ✗Requires local GPU/CPU resources
- ✗No cloud fallback (self-hosted only)
- ✗Older than Llama 4
Best Use Cases
💻 Local Development
Offline coding assistant
📱 Edge Apps
Mobile AI features
💬 Chatbots
Customer support
📧 Email Drafting
Quick responses
🔍 RAG
Document Q&A
🎓 Learning
AI experimentation
Ollama Command
ollama run llama3.2:11b
Available variants: llama3.2:1b, llama3.2:3b, llama3.2:11b, llama3.2-vision:11b