List models

Returns the full model catalog. Each model includes pricing, context window, capability flags, and its parameter schema in supported_parameters. This endpoint is public: a request without an API key is answered, and a request that sends an API key that does not verify (in Authorization: Bearer, x-api-key, or x-goog-api-key) returns 401.

Every model carries the OpenAI Model fields id, object, created (Unix seconds), and owned_by.

Every model carries a boolean native_inference field, true when EmpirioLabs serves the model on its own managed GPU infrastructure (the Native Inference designation in the catalog). Filter it together with category to select native models, for example native text models. Native models also name the exact Hugging Face repository being served in hugging_face_id.

pricing uses the OpenRouter convention of per-token USD strings. A model that is not priced per token, such as a video model priced per second or a speech model priced per 1,000 characters, reports "-1" in prompt, completion and input_cache_read. Its rates are listed in unit_pricing, one entry per unit and tier, and the full display breakdown is in pricing_rows.