This open-source AI model can be fine-tuned, distilled, and deployed anywhere. The latest instruction-tuned models are available in 8B, 70B, and 405B versions.

Models

• 405B flagship foundational model, supporting the broadest range of use cases.
• 70B high-performing, cost-effective model, supporting various use cases.
• 8B lightweight, ultra-fast model that can run anywhere.

Key Capabilities

• Tool usage
• Multilingual agent
• Complex reasoning
• Coding assistant

Make Llama Your Own

Build faster with our open ecosystem by choosing a range of differentiated product services to support your use case.

• Inference Choose between real-time inference or batch inference services. Download model weights to further optimize cost per token.
• Fine-tuning, Distillation, and Deployment Adapt and improve your applications with synthetic data, and deploy locally or in the cloud.
• RAG and Tool Usage Use Llama system components and extend the model with zero-shot tool usage and RAG to build agent-like models.
• Synthetic Data Generation Utilize 405B high-quality data to improve specialized models for specific use cases.

Quick Start with Partners

Model Evaluation

Measured on over 150 benchmark datasets covering multiple languages, with extensive human evaluations.

Model Pricing

As of 2024-07-23 12:00 PST, public pricing for hosting Llama 3.1 inference API is per million tokens. This table will be updated as more pricing information becomes available.

• Llama
• AI Blog - Learn AI from scratch