Gemini 3.7 Flash
Create an agent or open an existing agent in the builder, then follow the steps below. Model and feature availability may depend on your plan.
What it does
Google's Gemini 3.7 Flash joined the model selector — a multimodal workhorse with a one-million-token context window.
Before this
An addition to the selector rather than a new capability. Gemini 3.6 Flash had been the top Flash since July; 3.7 is the first one Google positions for complex coding and multi-step tool use rather than for cheap throughput, and it is the reason to move an agent up within the same family.
Why it matters
The balance of speed, cost, and context length to reach for on high-volume everyday agents.
The specs
Flash is Google's speed tier, and 3.7 is the first one they position for genuinely hard work — complex coding, multi-step tool use, and multimodal reasoning — rather than just cheap throughput.
- Context
- 1,048,576 tokens in, up to 65,536 out.
- Input
- Text and images.
- Actions
- Supported, including while reasoning.
- Reasoning
- Four levels in the Reasoning dropdown, from Fastest ($) to Smartest ($$$$).
- Plans
- Gold and Pro.
The million-token window is the reason to reach for it over a smaller model: a long transcript, a full knowledge base retrieval, and a lengthy conversation can all sit in context at once without you trimming the knowledge budget.