Together AI Raises $102.5M
Focused on scalability and efficiency, Together AI looks to scale their operations.
Created on November 29|Last edited on November 29
Comment
Together AI, a company committed to advancing open-source AI, recently announced that the company has secured $102.5 million in Series A funding, led by Kleiner Perkins, with substantial participation from NVIDIA, Emergence Capital, and others.
Performance Focused
Together AI is a company that specializes in providing high-performance AI services and products. Their offerings are designed to be extremely efficient and fast, enabling businesses and developers to handle AI tasks more effectively. A key aspect of their service is scalability; they provide solutions that can grow with a company, whether it's a startup just beginning to explore AI or a larger enterprise looking to expand its AI capabilities.
Fast Growth
Since its launch in June 2023, Together AI has experienced rapid adoption of its AI compute, training, and inference products among startups and enterprises. The company's goal is to expand its services and introduce new products that simplify the integration of AI into various applications. A key focus is on open-source AI, which is increasingly gaining traction due to its powerful generative models released regularly.
Together AI's platform enables developers to integrate these models or create their own through pre-training or fine-tuning. This flexibility, coupled with the company's industry-leading performance and reliability, has attracted a wide range of customers. These customers appreciate the freedom to own the results of their AI investments and the option to run their models on any platform.
Open Source Contributions
The company has taken a research-driven approach to building AI systems and products. Notably, it released the RedPajama-V2 dataset, the largest open dataset with 30 trillion tokens for training large language models (LLMs), which has seen extensive use and downloads. Additionally, the company's Chief Scientist, Tri Dao, and collaborators have developed FlashAttention v2, a tool widely used by major AI players to build leading LLMs.
The company's achievements in inference, utilizing techniques like Medusa and Flash-Decoding, have resulted in a highly efficient inference stack for transformer models. These developments are part of the Together Inference API, providing access to over 100 open models for quick inference.
The CEO of Together AI, Vipul Ved Prakash, emphasizes the belief that generative AI is a platform technology with the potential to significantly impact society. He stresses the importance of providing choices and options in the AI ecosystem, which will consist of both proprietary and open models. The company's mission is to empower researchers and developers to play an active role in shaping the future of AI.
The announcement:
Add a comment
Tags: ML News
Iterate on AI agents and models faster. Try Weights & Biases today.