Key Info

Perplexity has introduced the Orchestrator API, which lets developers pick the right model for each task while automatically selecting the serving provider based on availability, capacity, and latency. It unifies open-weight and proprietary models under one API with unified billing.

Highlights

  • Built on Perplexity's production model layer, which routes hundreds of millions of requests daily across 85+ models and more than a dozen providers.
  • Each model is billed at its published per-token rate, regardless of which provider serves it.
  • The API Console supports spend tracking by model, project, and API key.
  • Developers can set model allowlists and spend caps for cost control.