Google launched Gemini Pro 2.5 Experimental on the 25th of March. The model is at the time of writing dominating the AI benchmarks with its 63.8% SWE-Bench score and 18.8% performance on advanced reasoning tests. A good reason to tell you how to use this powerful model in your n8n automations!
Limitations of Gemini Pro 2.5 Experimental
Currently Google poses some strict limitations to the Gemini Pro 2.5 Experimental model in Google AI Studio. These limitations are at the time of writing:
- 2 requests per minute (RPM)
- 50 daily requests
Let me tell you how to use this model without any constraints.
Access to the model using OpenRouter
My method on accessing this model without running into the limits set by Google is to use OpenRouter. OpenRouter is an AI platform that gives users access to a wide varity of models without having to buy API credits for each platform seperatly.
The Gemini Pro 2.5 Experimental model is currently free to use at OpenRouter and the platform has recently gotten native integration with n8n!
Step by step access to Gemini Pro 2.5 Experimental
Here are the steps on how to get access to Gemini Pro 2.5 Experimental in OpenRouter.
- Go to OpenRouter
- Sign up if you haven’t got an account already
- Go to OpenRouter Chat
- Create a ‘New Room’
- Go to ‘Auto Router’ and hit the three dots beside it
- Click on the ‘Model’ dropdown and select the Gemini Pro 2.5 Experimental model
- Then click away (beside the overlay window) and chat with the latest Google model!
Use Gemini Pro 2.5 Experimental in n8n
You can not only use this powerful model by Google in OpenRouter Chat but also in n8n using the OpenRouter API.
Before you can use the OpenRouter API make sure you’ve generated an OpenRouter API key. You do this by following these steps:
- Go to OpenRouter
- Sign up if you haven’t got an account already
- In the top-right corner click on your profile picture and go to ‘Keys’
- In the API key settings click on ‘Create Key’. Save this key so this can later be used in n8n.
Got your OpenRouter API key? Let’s move on to n8n.
To use Gemini Pro 2.5 Experimental in n8n you have to make sure you’ve installed n8n. I recommend installing n8n using Docker Compose and Portainer.
If your n8n is ready to go, then follow these steps:
- Create a new workflow
- Add a node and click on ‘Advanced AI’
- Then choose ‘AI Agent’
