IBM Now Offers Meta’s Llama 3 Open Models on Watsonx

IBM has expanded its portfolio by offering Meta's Llama 3 open models on its Watsonx platform. This move is aimed at fostering next-generation enterprise-ready models.

To recall, in December of last year IBM and Meta launched the AI Alliance — a group of leading organizations across industry, startup, academia, research and government and it has since grown to more than 80 members and collaborators.

The Llama 3 models, which include pretrained and instruction fine-tuned language models with 8B and 70B parameter counts, are designed to support a wide range of use cases such as summarization, classification, information extraction, and content-grounded question and answering.

The Llama 3 8B model is optimized for faster training and deployment on edge devices, while the Llama 3 70B model is tailored for more demanding tasks like content creation, conversational AI, and language understanding¹. IBM's collaboration with Meta and the AI Alliance reflects a commitment to open innovation in AI, with the alliance now comprising over 80 members and collaborators.

IBM also hosts Code Llama 34B, a task-specific model for code generation and translation, on the Watsonx platform. These models are available for both SaaS and on-premises deployment, providing flexibility for clients to scale AI with their own data across a broader set of enterprise use cases.

For those interested in exploring these models, IBM offers a free trial on the Watsonx platform. It's an exciting development for AI builders and researchers who are looking to leverage these advanced models for various applications.