What it is
Together AI is a cloud platform for open models. Serverless inference serves chat, code, vision, image, video and speech models from families such as Qwen, DeepSeek, Kimi, GLM, Gemma, gpt-oss and Llama, with batch, provisioned throughput and dedicated deployments for heavier use. It also offers LoRA and full fine-tuning, code sandboxes, and GPU clusters with NVIDIA H100, H200, B200 and newer systems.
Everything is billed by usage: per million tokens for serverless models, per token processed for fine-tuning, with a minimum charge per job, and per GPU-hour for compute. It suits teams moving from closed APIs to open models or training their own. It is a developer service with no consumer app beyond a playground and Together Chat.
Best for
- Calling open models like Qwen, DeepSeek, Kimi and Llama via API
- Fine-tuning an open model on your own data
- Renting H100, H200 or B200 GPUs by the hour