OpenRouter — LLM Gateways Platform

Founded 2023|11-50 people|Free tier + pay-as-you-go (passthrough + 5.5% fee)

What is OpenRouter?

OpenRouter is a unified LLM gateway that routes requests to the best available provider for each model, with a single API key giving access to 300+ models from OpenAI, Anthropic, Google, Meta, Mistral, Cohere, and dozens of smaller providers. It exposes an OpenAI-compatible API, so any existing OpenAI SDK code works unchanged.

Two pricing tiers: a Free tier (25+ free-of-charge models, 50 requests/day, 20 RPM, raised to 1,000/day after $10+ in credits) and Pay-as-you-go (300+ models, passthrough provider rates, 5.5% platform fee on credit-card purchases / 5% on crypto). OpenRouter adds ~25-40ms latency over direct provider calls in typical conditions.

Major use cases: avoiding vendor lock-in across OpenAI/Anthropic/Google, fallback routing when a provider is down, cost optimization across price-equivalent models, and tracking spend per API key. Free credits expire after 365 days. As of 2026, OpenRouter is the most-used model gateway for AI startups and indie developers building model-agnostic applications.

Key features

Core capabilities this platform advertises.

300+ models across 60+ providers via one OpenAI-compatible API
Automatic routing to best available provider (latency, price, uptime)
Per-API-key spend tracking and rate limiting
Free tier: 25+ models, 50-1,000 req/day depending on credits
BYO API key for select providers (no markup)

Strengths and tradeoffs

What this tool does well, and the limitations to keep in mind.

Pros

Largest model catalog in the gateway space — 300+ models
OpenAI-compatible API drops in with zero code change
Reasonable platform fee — 5.5% over passthrough rates
Per-key spend tracking and isolated rate limits
Active free tier viable for prototyping

Cons

Adds ~25-40ms latency over direct provider calls
Credits expire after 365 days
5.5% fee + $0.80 minimum charge stings on small purchases
Free models can be flaky — timeouts and required cooldowns

Plans & pricing

What's included in each plan, and how the tiers compare.

Free Models

Forever

Access to 25+ free models
50 requests/day (1,000 with $10+ credits)
20 requests/minute
OpenAI-compatible API

Pay-as-you-go

Passthrough + 5.5%

300+ models, 60+ providers
Passthrough provider rates + 5.5% credit-card fee
5% fee on crypto purchases
Per-API-key spend tracking
Credits expire after 365 days

View official pricing page

Common use cases

Developers and teams who want one API for hundreds of LLMs without provider lock-in

Model-agnostic AI apps that swap providers freely
Fallback routing when a primary provider is degraded
Cost optimization across price-equivalent models
Side-by-side model comparison from a single API
Multi-tenant apps with per-tenant spend tracking

Using OpenRouter with Respan

OpenRouter and Respan enable multi-provider AI access with unified monitoring. Route requests across 290+ models with OpenRouter while tracking costs and performance with Respan.

Track costs across multiple LLM providers through OpenRouter unified gateway
Monitor routing decisions and fallback patterns to optimize provider selection
Compare performance across different models and providers using Respan analytics
Identify cost savings opportunities by analyzing provider-specific usage patterns

Monitor OpenRouter Requests with Respan