Skip to content
AI Model Radar

Gemini 3 Flash Preview

gemini-3-flash-preview
Google · chat · api · seen 9m ago
Preview Alert me on changes
Context
1M
Max output
65.5K
Input $/1M
$0.50
Output $/1M
$3.00
Modalities
text
Released
17 Dec 2025
AI summary
● machine-written

Google releases Gemini 3 Flash Preview with 1M context window

Gemini 3 Flash Preview is a text-based chat model in preview status from Google, designed for agentic workflows and multi-turn interactions. It supports a 1 million token context window with up to 65.5K token output, priced at $0.50 per million input tokens and $3.00 per million output tokens. The model includes reasoning capabilities and supports tools, caching, and code execution.

What's new
  • Released December 17, 2025
  • Supports 1M token context window with 65.5K max output
  • Includes thinking/reasoning mode with higher benchmark performance (37.8 on Intelligence, 97.0 on Math)
  • Available via Google API with Batch API and flexible inference support
  • Supports caching, code execution, computer use, and search grounding
Best for
Agentic workflows and multi-turn chatCoding assistance and code execution tasksLong-context reasoning with thinking modeCost-conscious applications requiring moderate performance
Sources

Source: https://ai.google.dev/gemini-api/docs/models/gemini-3-flash-preview