Nebius runs open models in production through Nebius Token Factory, the current name on its site for the inference product that was previously marketed as AI Studio. It offers access to leading open text, code and reasoning models, inference tuned for latency, throughput and cost, scalable infrastructure for sustained traffic, and post-training to adapt models with your own data.
You can start free with serverless inference through an API, or choose dedicated endpoints for isolated capacity. The API is OpenAI-compatible.
See Nebius's Token Factory pricing page for rates.