OpenRouter guide

Use OpenRouter's free routes without guessing.

Set up a key, choose the right route, understand its limits, and make a safe first request.

Fact checked July 15, 2026

Four deliberate steps

OpenRouter provides the API. Free Model observes and explains selected routes; your requests go directly to OpenRouter.

01

Create a dedicated key

Sign in to OpenRouter and create an API key. Keep it in a server-side environment variable; never paste it into client code or a public repository.

02

Choose a route deliberately

Use an exact model ID ending in :free when identity matters. Use openrouter/free when a compatible model selected from the current free pool is acceptable.

03

Plan around the quota

Free variants have account-level request limits. Treat 429 responses as temporary, honor retry guidance, and do not create extra keys to bypass limits.

04

Verify the response and cost

Inspect the returned model, usage, and your OpenRouter account. A successful response proves that request worked; it does not guarantee future capacity.

Open OpenRouter API keys

OpenRouter officially offers account signup and API key creation. Free Model reports new-user eligibility as independently verified only after a recent, ordinary new account completes that flow. A catalog statement by itself is labeled as not independently tested.

Exact model or free router?

Specific route

publisher/model:free

Choose this when you need a known model, context window, or capability. Availability can still vary because the free upstream has finite capacity.

Pool router

openrouter/free

OpenRouter selects a compatible model from its current free pool. The response identifies the model actually used. Do not describe this router as a foundation model.

Need reproducible model behavior? Prefer an exact route. Need a low-friction experiment and can accept a changing model? The free router may be the better fit.

Current observed routes

Make the smallest useful call

Start with a non-sensitive prompt and a short response. The same chat-completions endpoint works with an exact free model ID. Replace only the model value.

Terminal · cURLOPENROUTER_API_KEY
curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openrouter/free",
    "messages": [
      {"role": "user", "content": "Reply with one short sentence."}
    ]
  }'

The model ID is part of the billing and routing decision. Check it before sending a request, especially if your client enables fallbacks.

The account cohort matters

OpenRouter currently documents the following platform limits for model IDs ending in :free. These are account-level limits, not promises of upstream capacity.

Credits purchased, all timePer minutePer day
Less than $102050
At least $10201,000
  • Multiple API keys or accounts do not increase the documented global capacity.
  • A negative account balance or per-key credit cap can produce a 402 even on a free model.
  • A 429 can come from OpenRouter's account limit or from upstream capacity. Retry with backoff.
  • Free Model keeps results from different account cohorts separate instead of blending them into one status.

Check the primary source

Our pages summarize current facts, but OpenRouter remains authoritative for its API and account.

Browse observed routes