
Together AI
together.ai · Inference & Hosting
Cloud for running, fine tuning and training open models, plus on demand GPU clusters.
Together AI is a cloud for open models. Its serverless inference API serves a large catalog of open models (chat, code, vision, image and embeddings) behind an OpenAI compatible endpoint, billed per token, with batch inference for offline jobs. It also offers dedicated model and container inference, provisioned throughput with SLAs, managed fine tuning and post training (SFT and DPO), and GPU clusters with H100, H200, B200 and GB200 class hardware.
Prices differ by model and change often, so check the official pricing page. For agent builders it is available as a provider in the Vercel AI SDK (@ai-sdk/togetherai), in LiteLLM and through OpenRouter. Alternatives include Fireworks AI, Groq, Baseten and Replicate, or self hosting with vLLM.
Together AI is one of 10 inference and hosting tools on Drydock. See all inference and hosting tools or how we check facts.