The Gemini API improvements include simpler controls over thinking, more granular control over multimodal vision processing, and ‘thought signatures’ to improve function calling and image generation.