Multimodal models from Google
Gemini models for text, images, and long context, with a generous free tier for prototyping.