Meta Llama

Meta Llama provides a suite of open-source large language models and multimodal AI models, including Llama 4 Scout and Llama 4 Maverick. It offers an API for developers to integrate these models into their applications, focusing on performance, efficiency, and safety. The platform supports various AI capabilities such as conversational AI, image understanding, and code generation.

Meta Llama directly offers large language models (LLMs) and multimodal models (Llama 4 Scout, Llama 4 Maverick) that support conversational AI and image understanding. They provide API access for developers to integrate these models into their applications. The website explicitly mentions 'Llama Protections' and 'Llama Guard' models, indicating a focus on safety and alignment. While not explicitly called 'fine-tuning', the ability to 'adapt' models and the provision of different model sizes (e.g., 1B, 3B, 70B, 400B) suggest customization capabilities. The 'Llama API' and 'Llama Stack' are geared towards enterprise deployment and scaling. The site also highlights 'Code specific version of Llama 2' for code generation and mentions 'Research & Publications' and 'Llama Defenders Program' which aligns with security and red teaming.