OpenRouter
OpenRouter is a unified API and interface for accessing 600+ AI models from 80+ providers. Developers can compare models and prices, route requests automatically, and pay for usage with credits, with a free tier of 25+ models.
About OpenRouter
Overview
OpenRouter is a unified interface for accessing AI models. According to its website, it offers 600+ active models from 80+ providers through a single, OpenAI-compatible API. Developers can create an account, buy credits, generate an API key and start making requests.
Key Features
- Large model catalog: 600+ models from 80+ providers, including models from Google, OpenAI and Anthropic.
- OpenAI-compatible API: Create an API key and send requests in an OpenAI-compatible format.
- Model and price discovery: Find and compare models and prices for your prompts.
- Credits: Buy credits that can be used with any model or provider.
- Bring your own keys (BYOK): Connect your own provider keys so the provider bills you directly for inference.
- Team and enterprise options: Business and Enterprise plans with organization features, support and invoicing options.
Use Cases
- Building apps that need access to many AI models through one API.
- Comparing model quality and pricing for specific prompts.
- Routing requests across providers without managing separate integrations.
- Running AI agents and coding workflows, as described in the site's tutorials and blog posts.
Pricing
OpenRouter states that there are no subscriptions and that you pay for what you use. A free tier includes 25+ free models from 4 free providers with 50 requests per day. Standard and Business plans work by buying credits, with a fee of 5.5% on Standard credit purchases and 8% on Business credit purchases. Enterprise pricing is set by contract based on volume, prepaid credits and annual commitments, with invoicing. BYOK includes $25,000 of list-price inference per month with no OpenRouter fee on Standard and Business, with a 5% fee above that allowance.
Who It Is For
OpenRouter is aimed at developers, indie hackers, AI-native startups, growing teams and enterprises that want one interface for many AI models and providers.