Note: you need the Open AI Chat-compatible endpoint URL (see below)
- Generate an API key here https://ai.google.dev/gemini-api/docs/api-key
- Set up BYOM on Leo
- Label: whatever you want
- Model request name: gemini-2.5-pro-exp-03-25
- Server endpoint: https://generativelanguage.googleapis.com/v1beta/openai/chat/completions
- API key: obvious
I left context size at the default value.
Other Gemini models can be used but the one I used above is the latest 2.5 that can be used on the free tier.
Other models are here: https://ai.google.dev/gemini-api/docs/models