← All models

Gemini 3 Pro

Published Publisher Googlegemini-3-pro-preview

Overview

Gemini 3 Pro is Google’s preview reasoning model for complex multimodal understanding, coding, and agentic workflows. It was released on November 18, 2025. (blog.google)

Capabilities and uses

  • Accepts text, images, video, audio, and PDFs; generates text. The documented limits are 1,048,576 input tokens and 65,536 output tokens.
  • Supports thinking, function calling, code execution, structured output, Google Search grounding, URL context, context caching, and the Batch API.
  • Intended uses include software development, multimodal document and media analysis, and applications that use tools or carry out multi-step tasks. (ai.google.dev)

Limitations and availability

  • The preview endpoint was shut down on March 9, 2026 and is no longer available through the Gemini API. Google recommends gemini-3.1-pro-preview as its replacement.
  • This model does not generate images or audio and does not support the Live API or computer use. (ai.google.dev)

Pricing

At launch, Google announced preview pricing of USD $2 per 1 million input tokens and USD $12 per 1 million output tokens for prompts of 200,000 tokens or fewer, through the Gemini API in Google AI Studio and Vertex AI. These are historical prices, not an available current offer. Pricing for larger prompts was not verified for this specific model. The current official pricing page no longer lists this shut-down model. Checked November 17, 2026. (blog.google)

Official sources