Google's open multimodal model with 128K context and 140+ language support.
Add a second model to compare.