IBM has launched Granite 3.0, the latest version of its large language models (LLMs) tailored for enterprise use, blending high performance with cost-efficiency. These models, built with over 12 trillion tokens, support 12 natural languages and 116 programming languages, making them ideal for diverse industries like cybersecurity and text generation. Granite 3.0’s versatility allows businesses to customize AI solutions to their specific needs, enhancing their operational capabilities.
A key aspect of Granite 3.0 is IBM’s dedication to open-source principles. All models are released under the Apache 2.0 license, ensuring transparency by openly sharing training data and methodologies. This approach aims to foster trust in AI and encourage collaborative innovation across the tech community.
Granite 3.0 is now accessible through platforms like IBM’s watsonx, Google Vertex AI, and Nvidia, making it easier for businesses to integrate these advanced AI capabilities. Future updates will include expanded language support and new features, such as image-based input, broadening the models’ applicability even further.
Reflecting IBM’s commitment to sustainable technology, Granite 3.0 was trained on Blue Vela, a system powered entirely by renewable energy. This eco-friendly initiative underscores the company’s focus on delivering powerful AI tools while minimizing its environmental impact, solidifying Granite 3.0’s standing as a top choice for enterprises looking to harness AI responsibly