Unlocking Gemini 3 Pro in Your Terminal: A Deep Dive into the New Gemini CLI

Google is progressively integrating its powerful new Gemini 3 model across its suite of developer tools. Following its introduction in platforms like Stitch AI, the state-of-the-art AI is now making its way into the Gemini Command Line Interface (CLI), bringing a new level of intelligence and efficiency directly to your terminal. Let’s explore the exciting new updates and how you can get started with Gemini 3 Pro on the Gemini CLI.

Gaining Access to Gemini 3 Pro

[00:41:645] [Gemini 3 Pro on Gemini CLI Waitlist Documentation]

Currently, using Gemini 3 directly within the CLI requires joining a waitlist. The process is straightforward, and Google is processing applications quickly, with many users gaining access within a day or two. It’s a highly responsive system designed to get developers on board with the latest features as soon as possible. If you are already using the Gemini API with your own key, you may be able to bypass the waitlist and start using the model immediately.

We’re excited to bring Gemini 3 Pro to Gemini CLI. For Google AI Ultra users (Google AI Ultra for Business is not currently supported) and paid Gemini and Vertex API key holders, Gemini 3 Pro is already available and ready to enable. For everyone else, we’re gradually expanding access through a waitlist.

A Polished, Seamless Terminal Experience

One of the most significant recent updates to the Gemini CLI is a complete overhaul of its user interface. The team has rebuilt the foundation to create a smoother, more intuitive, and visually stable environment that feels less like a traditional terminal and more like a polished graphical application.

[01:57:045] [Google for Developers Blog Post about the new Gemini CLI UI]

This new rendering foundation eliminates common terminal annoyances like flickering screens and bouncing input prompts. Key enhancements include:

• Mouse-Based Navigation: You can now use your mouse to click and navigate directly within the input prompt, making it easier to edit long commands and copy-paste text.
• Sticky Headers: When you’re scrolling through long outputs or file edits, important context like filenames remains “stuck” at the top of the view, ensuring you never lose track of what you’re working on.
• Flicker-Free Resizing: The display is now clean and polished, even when you resize the terminal window, removing visual glitches and artifacts.
• Stable Input Prompt: The input field is now firmly anchored to the bottom of the terminal, providing a consistent and predictable user experience.

How to Enable and Use Gemini 3 Pro in the CLI

Before you can use Gemini 3, ensure your CLI is updated to the latest version.

[01:28:145] [npm install command for Gemini CLI]

Run the following command in your terminal to install or update the package:

npm install -g @google/gemini-cli

Once you’ve been granted access from the waitlist and have updated the CLI, you need to enable the preview features.

[04:29:945] [Gemini CLI settings command in the terminal]

1. Run the /settings command inside the Gemini CLI.
2. Navigate to Preview Features and set it to true. This action unlocks the ability to use the newest models.

[04:47:745] [Selecting the Gemini 3 Pro model in the Gemini CLI]

With preview features enabled, you can now select your desired model. Run the /model command to open a dropdown menu. For the most powerful experience, select the Pro (gemini-3-pro-preview) option. The CLI will attempt to use this model first before falling back to older versions like gemini-2.5-pro if necessary, especially during long-running tasks or high server load.

Monitoring Performance and Usage

[05:54:145] [Model statistics for nerds in Gemini CLI]

If you’re curious about performance or which models are being used for your tasks, you can get detailed statistics. Run the /stats model command to see a breakdown of API requests, average latency, errors, and token usage for each model, giving you valuable insight into your workflow.

Putting Gemini 3 CLI to the Test

To see its capabilities, the model was tested on several benchmark tasks.

[06:07:345] [Benchmark prompt for creating a movie tracker app]

A prompt to build a good-looking movie tracker app with an Expo framework yielded impressive results. The CLI quickly scaffolded the entire application, including setting up a custom theme, creating components, handling API calls, and managing local storage.

[07:12:945] [iOS style terminal calculator created by Gemini CLI]

Other tasks, like creating a Go-based terminal calculator and an offline-first Svelte Kanban app, were also executed flawlessly, scoring a perfect 100/100 on evaluation benchmarks. The CLI’s ability to handle complex requirements, from UI aesthetics to backend logic, showcases a significant leap in agentic coding capabilities.

[08:20:045] [Benchmark results for Gemini 3 Pro on Gemini CLI]

While some complex tasks still faced challenges, often due to the model falling back to Gemini 2.5 Pro, the overall performance is strong. The Gemini 3 Pro integration, combined with the major UI enhancements, makes the Gemini CLI a powerful and enjoyable tool for any developer looking to leverage AI in their terminal workflow.