Arcee AI has launched SuperNova, a 70 billion parameter language model designed for enterprise use, emphasizing data privacy, model stability, and customization. Built on Meta’s Llama-3.1-70B-Instruct architecture, SuperNova can be deployed directly within a company’s own infrastructure, starting with AWS and soon expanding to other platforms. Notable features include advanced instruction-following capabilities, the ability to fine-tune and retrain within the enterprise environment, and several open-source components to aid developers. The model claims superior performance in areas such as mathematical reasoning and emphasizes control over AI assets without reliance on external APIs.
Table of contents
Technical innovationsEnterprise deployment and customizationCustomization and continuous improvementOpen source componentsPerformance claims and benchmarksImplications for Enterprise AI strategyThe AI sovereignty dilemmaSort: