Together AI

Together AI is a cloud-based platform for creating open-source generative AI and infrastructure for AI model development. The company advances AI by providing innovative open-source research, models, and datasets. Its decentralized cloud services enable developers and researchers to train, refine, and deploy generative AI models across organizations of all sizes.

The company's flagship platform, Together API, facilitates the training, fine-tuning, and inference of open-source large language models (LLMs) and custom models. Customers can integrate leading open-source LLMs like Llama-2, Falcon, Vicuna, and Mistral/Mixtral, or bring their own models.

Its product offering includes Together Inference, an inference stack to run open source models, Together Finetuning, a platform to customize open-source models with private data, Together custom models, to build models from scratch, and Together GPU clusters which offers compute clusters for training and fine-tuning.

The company also offers consulting services to design customized AI models for specific use cases. Together AI's research initiatives include RedPajama for open-source chat models and Evo, a biological foundational model for understanding and generating genomic sequences.

In December 2024, the company acquired development environment infrastructure provider CodeSandbox. The company aimed to integrate CodeSandbox's technology into its Together Inference Platform, creating the first platform with built-in code interpretation.

Key customers and partnerships

Together AI's customers include AI startups like Nexusflow.ai, Voyage.ai, and Pika Labs, as well as enterprises from various industries.

The company has notable partnerships with Snowflake, to bring Arctic LLM to Enterprise customers, and Meta, to release Meta Llama 3 for inference and fine-tuning.