gemini-3.1-pro-preview model and pricing

A third-generation Pro preview model for multimodal understanding, agents and coding. Accepts text, images, video, audio and PDFs; produces text, not audio. Context: 1,048,576 tokens; maximum output: 65,536. For prompts up to 200K tokens, input/output/cache-read rates are $2/$12/$0.20; above that threshold they are $4/$18/$0.40. Supports function calling, structured output, code execution, file search, search grounding and context caching.