Free Model logoFree Model

Guide

Set up a free AI model.

Choose a model, connect it, and test one request. OpenRouter is supported now, and more free model platforms will be added over time.

Fact checked July 18, 2026

Step 1

Choose a free AI model

Choose the model you want to test and use in the examples below.

Step 2

Create and add your API key

The models currently available here use OpenRouter. Create an account if needed, then generate and paste an API key.

Create accountCreate API key

Stored only in your browserYour key is used only in this browser and sent directly to OpenRouter. Free Model never receives, stores, or logs it on our servers.

Step 3

Run one test request (Optional)

Send one message in this browser to confirm that the selected model and API key work.

Quick test (Optional)

Send a message to confirm that this model and your API key work.

Step 4

Use it in your project

Choose your language, copy the example, and use the colored Base URL, API key, and Model ID in your project.

curl 'https://openrouter.ai/api/v1/chat/completions' \
  -H "Authorization: Bearer ${OPENROUTER_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{"model":"tencent/hy3:free","messages":[{"role":"user","content":"Reply with exactly: free-model-ok"}],"max_completion_tokens":16}'

If it fails

Match what you see with the fix below.

  1. 01

    Authentication failed

    What you see
    The request returns 401, Missing Authentication, or Invalid API key.
    How to fix it
    Return to Step 2, paste a valid OpenRouter API key, and run the test again. If the error remains, create a new key and replace the old one.
  2. 02

    The model does not reply

    What you see
    The test shows a model error, provider error, or finishes without an AI response.
    How to fix it
    Confirm the selected Model ID, retry once, then choose another free model from the dropdown if the provider is still unavailable.
  3. 03

    The free route is busy

    What you see
    The request returns 429 or a rate-limit message.
    How to fix it
    Wait a short time and retry. If you need an immediate response, select a different free model and test that route instead.