Cloud platform for fast, low-cost inference and fine-tuning of open models.
Together AI provides serverless and dedicated inference for hundreds of open-source models (Llama, Qwen, DeepSeek, FLUX, and more) via an OpenAI-compatible API, plus fine-tuning and GPU clusters. Pricing is usage-based per token or per GPU-hour, with no free consumer tier beyond initial credits.
Usage-based per-token pricing for serverless inference; per-GPU-hour for dedicated endpoints and clusters. Sign-up credits available.