Add a free model as a custom OpenAI-compatible model in Cursor.
Cursor lets you add custom models under Settings → Models → Add Model, as long as the provider exposes an OpenAI-compatible /chat/completions endpoint.
- Open Settings → Models
- Click Add Model
- Fill in:
Model name: llama-3.3-70b-versatile
Base URL: https://api.groq.com/openai/v1
API key: your-groq-key # get at console.groq.com/keys
- Disable "OpenAI API Key" verification if Cursor asks — it only applies to official OpenAI models.
| Provider | Base URL | Get Key | Credit Card? |
|---|---|---|---|
| Groq | https://api.groq.com/openai/v1 |
console.groq.com/keys | No |
| NVIDIA NIM | https://integrate.api.nvidia.com/v1 |
build.nvidia.com | Phone verification |
| OpenRouter | https://openrouter.ai/api/v1 |
openrouter.ai/keys | No |
| Cerebras | https://api.cerebras.ai/v1 |
cloud.cerebras.ai | No |
| Mistral (La Plateforme) | https://api.mistral.ai/v1 |
console.mistral.ai/api-keys | Phone verification |
- Cursor's Tab/Composer features are tuned for specific models and may work best with the officially supported ones — free models work great for chat and inline edits, results vary for agentic multi-file edits.
- Rate limits apply per provider — see the Provider Directory for details.
See the full Provider Directory and Quick Reference in the main README for all 33 free providers, or browse free-llm.com.