Together AI

About Together AI

Together AI was founded in 2022 with the thesis that the future of AI is built on open-source models, and that developers need infrastructure optimised for running, customising, and deploying those models at scale. The company has built a managed inference platform that serves over 100 open-source language models, embedding models, and image generation models via a unified API. The Together AI platform covers several key use cases for AI developers. The Inference API provides access to models including Llama 3, Mistral, Mixtral, Code Llama, DBRX, Qwen, and many others via an OpenAI-compatible endpoint, enabling straightforward migration from proprietary to open-source model providers. The Fine-Tuning service enables organisations to customise base open-source models on their own data, with Together managing the GPU infrastructure and returning a deployable fine-tuned model. Embeddings and image generation APIs round out the platform. Together has positioned itself as the destination for open-source AI infrastructure, and has invested heavily in inference optimisation—competing with Groq on speed for GPU-based inference. The company has raised over $100 million in venture funding and has notable traction with developer-focused companies and AI startups. For European businesses, Together AI presents a mixed compliance picture. The company is incorporated in the United States (Delaware) and processes data on US infrastructure. There is no EU data residency option as of early 2026. Together publishes a GDPR privacy addendum and offers data processing agreements for enterprise customers, but the absence of EU data centres and independent security certifications (ISO 27001, SOC 2 Type II) means the platform requires careful evaluation for regulated EU use cases. Together's explicit policy is that customer data and fine-tuning datasets are not used to train shared models, which is an important baseline for privacy, but does not address the fundamental data residency concern.