Meta

The flagship model in Meta's Llama 3.3 family with a 128,000-token context window, designed for enterprise and research applications. It delivers state-of-the-art performance across reasoning, coding, and language understanding, with enhanced multilingual support and robust safety guardrails. Llama 3.3 70B excels in complex tasks like detailed content creation, technical problem-solving, and advanced applications requiring sophisticated reasoning.

An efficient open-weight model from Meta's Llama 3.1 family with an 8,000-token context window, optimized for accessibility and local deployment. It provides strong performance with improved multilingual support and instruction following, ideal for cost-effective solutions like lightweight chatbots, educational tools, and on-device AI applications.

A sophisticated multimodal mixture-of-experts model from Meta with a 10,000,000-token context window, offering unprecedented document processing capabilities. It excels in multimodal understanding, processing text and images with exceptional reasoning and knowledge. As Meta's most advanced model, it delivers superior performance for enterprise applications, research, and complex analytical tasks.

A balanced multimodal mixture-of-experts model from Meta with a 10,000,000-token context window, offering comprehensive document processing capabilities. It provides strong multimodal understanding of text and images with 17 billion active parameters across 16 expert modules. This model delivers excellent performance for general applications requiring robust reasoning and image analysis.

A mid-sized model from Meta's Llama 3.2 family with a 128,000-token context window, offering improved reasoning and instruction following. It provides strong performance across general tasks with efficient resource utilization, making it suitable for applications requiring a balance of capability and deployment efficiency.

A highly compact model from Meta's Llama 3.2 family with a 32,000-token context window, designed for resource-constrained environments. It offers surprisingly strong performance for its size, making it ideal for edge devices, on-device applications, and scenarios requiring minimal computational resources.

A lightweight model from Meta's Llama 3.2 family with a 64,000-token context window, offering impressive capabilities for its compact size. It balances efficiency and performance for applications requiring local deployment or constrained resources while maintaining reasonable context handling and reasoning abilities.

The largest model in Meta's Llama 3.2 family with a 128,000-token context window, offering exceptional reasoning, instruction following, and knowledge encoding. It delivers state-of-the-art performance for complex tasks requiring sophisticated analysis, detailed content generation, and technical problem-solving, making it suitable for research and enterprise applications.