It's a marketplace of inference providers. Inference is when you run the llm and let people send requests to it. The big models: Gemini, gpt and Claude have development and inference provided by their parent companies (Google, open ai and anthropic). The open models have a multitude of providers....