Perplexity's Gateway API unifies access to top AI models
Unified API for Anthropic, OpenAI, Google, xAI, and Perplexity models now live
Perplexity AI has officially launched its Gateway API, a unified interface designed to simplify access to leading AI models from multiple providers. The API consolidates models from Anthropic (Claude Opus 5), OpenAI (GPT-5.6 family), Google (Gemini 3.5 Flash), xAI (Grok 4.5), and Perplexity itself through a single endpoint, eliminating the need to manage separate integrations. This approach uses OpenAI Chat Completions and Anthropic Messages formats for compatibility, while automatically routing requests to the healthiest model deployment and handling failover scenarios seamlessly.
Pricing follows a transparent per-token model aligned with each provider’s published rates, with no additional per-request fees. The Gateway API now supports cutting-edge models like Claude Opus 5, the GPT-5.6 family (Sol, Terra, Luna), Moonshot AI’s Kimi K3, and others, making it a practical choice for developers needing to switch models without rewriting code.
- Unified API supports Anthropic, OpenAI, Google, xAI, and Perplexity models via a single endpoint
- Per-token pricing at published rates with no per-request fees; automatic failover included
- Supports new models: Claude Opus 5, GPT-5.6 family, Grok 4.5, and Moonshot AI’s Kimi K3
Why It Matters
Developers can now switch AI models without rewriting code, reducing integration complexity and accelerating deployment.