- Price
- Provider price plus a 5.5% platform fee
- Free plan
- Yes
- Setup
- Needs a developer
- Up and running
- About an hour if your app already calls OpenAI; an afternoon from scratch.
What OpenRouter is for
Instead of opening an account, adding a card and learning the API at every AI company (OpenAI, Anthropic, Google...), you use one key and switch models by changing a name in your code. There's one account, one credit balance and one bill, with the same per-token price each provider charges.
How it plugs into your product
- 1Create an OpenRouter account; you can already test the free models, up to 50 requests a day.
- 2Buy credits by card; the 5.5% platform fee is charged on top of each purchase.
- 3Create an API key and point your existing OpenAI-style code to OpenRouter, keeping the key on the server.
- 4Pick the model by name (for example a Claude, GPT or Gemini model) and set provider and privacy preferences.
Does it fit your idea?
It fits if
- You're still deciding between Claude, GPT and Gemini and want to compare them on real users without three accounts.
- Your app lets the customer choose the AI model, like a writing tool with a model picker.
- You want a backup: if one provider goes down, requests fall back to another that serves the same model.
Look elsewhere if
- You spend thousands of dollars a month on one model: the 5.5% fee on every credit purchase adds up; go direct.
- Your contract or compliance team requires a direct agreement with the model maker.
- You plan to run the product on free models: 50 requests a day (1,000 after buying $10) won't hold real users.
What people only find out after building
Read this before you build on it. Each item links to the official page that says so.
The 5.5% fee is charged when you buy credits and is non-refundable. You can only refund unused credits within 24 hours of the purchase, and crypto payments never.
Official source: openrouter.aiOpenRouter reserves the right to expire unused credits one year after purchase.
Official source: openrouter.aiBy default, the same open model can be served by different providers, picked mostly by price, and providers may run it at different precision (quantization). Answers can vary between calls unless you pin providers.
Official source: openrouter.aiRouting allows providers that collect data by default. If you handle customer data, set data_collection to "deny" so only providers that don't keep it are used.
Official source: openrouter.aiFree models are capped at 20 requests a minute and 50 a day (1,000 after $10 in credits). Failed attempts count, and a negative balance blocks them.
Official source: openrouter.ai
Price
Provider price plus a 5.5% platform fee
25+ free models, 50 requests a day (1,000 after buying $10 in credits).
- Price per million tokens (main model)
- Same as the provider + 5.5% fee
- Models
- 500+ models, 80+ providers
- Free tier
- 50 requests/day on free models
Checked on Oct 28, 2026 · official pricing
What it costs in practice
Estimate: 1,000 users × 30 chats a month × about 2,000 tokens per chat (1,500 in, 500 out) = 45M input and 15M output tokens. On Claude Sonnet 5.5 at Anthropic's list price ($2 / $10 per million) that's about $240 in usage, plus 5.5% on the credits you buy: about $253/month. On a cheap model like gpt-6-luna it's about $12 + fee ≈ $13.
Estimate from the official price. Check it before you decide.
At a glance
- Free plan: yes
- Open source: no
“Not stated” means the official site doesn’t say.
Alternatives
Compare side by side
Faundr
Picked your tools? Build it with AI, safely.
Faundr keeps your project’s memory, blocks what can’t be undone and tells you if it’s safe to launch, inside Claude Code and Codex.