Together AI

The product is a full-stack AI cloud platform that provides inference APIs, model fine-tuning and custom training, and GPU clusters and storage for running and adapting open-source models. It addresses the need to deploy, scale, and customize AI models without building underlying infrastructure. It sells to developers, data teams, and IT teams in businesses building AI applications. It is delivered as cloud infrastructure and APIs with serverless, batch, and dedicated inference options and on-demand GPU clusters.

Key features

• Serverless inference via APIs
• Batch inference for workloads
• Provisioned throughput with SLAs
• Dedicated model inference
• Dedicated container inference
• GPU clusters at scale
• AI Factory custom infrastructure
• Developer Environments Sandbox
• Managed storage for weights and data
• Custom training with reinforcement learning
• Fine-tuning with your data
• Model evaluations and measurement