Arcee AI specializes in developing and providing platforms for building and training Small Language Models (SLMs) within secure environments. Here are the key aspects of what the company does:

Core Offerings

End-to-End Platform: Arcee AI offers a comprehensive platform that enables organizations to build, train, and deploy SLMs tailored to their specific needs. This platform operates within a virtual private cloud, ensuring high security and data privacy.

Model Merging: This innovative technique allows users to combine multiple AI models into a single, more efficient model without increasing its size or requiring additional training. This enhances performance while minimizing computational costs.

Spectrum: Another proprietary technology that optimizes the training of language models by focusing on specific useful layers, significantly improving efficiency and reducing resource requirements.

Arcee Cloud: A hosted Software as a Service (SaaS) solution that simplifies the process of training and deploying SLMs, making it accessible to users with varying technical expertise.

Enterprise Solutions: Arcee AI provides customized solutions for highly regulated industries like healthcare, finance, and legal services, allowing organizations to maintain complete control over their data and models.

Recent Developments

Arcee AI recently secured $24 million in Series A funding, which will support its mission to democratize access to SLMs and enhance their capabilities for various business applications. The company aims to address the growing demand for efficient and secure generative AI solutions across diverse sectors.

Products of Arcee AI

• Arcee Cloud: A hosted Software as a Service (SaaS) platform that allows users to train and deploy SLMs with ease, even for those with limited technical expertise.

• Model Merging: A proprietary technique that enables users to combine multiple language models into a single, more efficient model without additional training or increased size.

• Spectrum: This method optimizes the training of language models by focusing on specific useful layers, improving efficiency, and addressing issues related to model retention of previous knowledge.

• Arcee Spark: A powerful 7 billion parameter language model that excels in performance for its size, suitable for various applications such as chatbots and customer service automation.

• Arcee Agent: Another 7 billion parameter model specifically designed for function calling and tool use, allowing for intelligent middleware capabilities and real-time applications.

• Llama Spark: A conversational model developed for engaging in natural dialogues and informative conversations.

• Arcee Scribe: A versatile chat model tailored for various conversational tasks.

• Arcee Lite: A compact yet effective 1.5 billion parameter model aimed at providing efficient AI capabilities in resource-constrained environments.

• Arcee Nova: A state-of-the-art 72 billion parameter model designed for high-performance applications.