IBM launches Granite 3.1 AI models with advanced features for enterprises

IBM has launched its Granite 3.1 Large Language Models, which are open-source and designed for enterprise use. This new version features a 128K token context window, significantly larger than previous models, allowing for better text processing and analysis. The Granite 3.1 models outperform competitors like Google and Meta on key benchmarks. They include both dense models and Mixture of Experts variants, trained on over 12 trillion tokens, making them suitable for various business applications, including code generation and translation. Additionally, Granite 3.1 introduces image-in/text-out functionality and improved language support for twelve languages. The models are available on IBM's Watsonx platform and other cloud services, promoting wider adoption of AI in business settings.