Type: url
Original: https://developers.googleblog.com/new-gemini-api-updates-for-gemini-3/
Status: generated
Language: en
Summary
Google’s developer blog post announcing the Gemini API changes that accompanied the release of Gemini 3: a thinking_level parameter for controlling the depth of the model’s reasoning, a media_resolution parameter for image, video and document inputs, enforced thought signatures for function calling and image generation, the ability to combine Grounding with Google Search and URL context with structured outputs, and a usage-based price for Search grounding. It closes with prompting best practices for Gemini 3 Pro, such as keeping temperature at its default of 1.0 and placing instructions after long context.