Gemini 3 Pro is now integrated directly into the Gemini CLI, giving developers access to Google’s most advanced reasoning model right inside the terminal. With agentic coding, multimodal understanding, better tool use, and state-of-the-art command execution, the Gemini CLI becomes one of the most powerful developer tools available.
This tutorial walks you through everything you need to start using Gemini 3 Pro in Gemini CLI — including installation, setup, features, and practical examples.
Currently, Gemini 3 Pro is only available to Google AI Ultra subscribers and paid Gemini API users. But there’s good news — as an all-in-one AI platform, GlobalGPT has already integrated Gemini 3 Pro, and you can try it for free.
What You Need Before You Start
To use Gemini 3 Pro in Gemini CLI, you must meet one of the following eligibility conditions:
Eligible Users
- Google AI Ultra subscribers (CLI access available now)
- Developers with a paid Gemini API key
- Gemini Code Assist Impresa (coming soon)
Not Yet Eligible (Waitlist Required)
- Google AI Pro
- Gemini Code Assist Standard
- Free-tier Gemini users
1. Install or Update Gemini CLI
Use npm to install or update to the latest version:
npm install -g @google/gemini-cli@latest
Then confirm your version:
gemini --version
You should see version 0.16.x or higher.
2. Enable Gemini 3 Pro in the CLI
Once installed, run:
/settings
Then toggle:
Preview features → ON
After this, Gemini CLI will default to Gemini 3 Pro.
3. Use Gemini CLI for Intelligent Coding
Gemini 3 Pro supports agentic coding, meaning it can analyze instructions, generate full multi-file scaffolds, and produce deployable codebases.
Example: Generate a ready-to-run 3D web app
Prompt:
Build a photorealistic 3D voxel simulation of the Golden Gate Bridge using Three.js.
Include lighting controls, fog slider, water shader, traffic simulation, and produce
a single HTML file named golden_gate_bridge.html.
Gemini CLI will:
- create an execution plan
- build the entire HTML/JS project
- optimize Three.js imports
- output a complete runnable file
4. Turn a Sketch Into Real UI
Gemini 3 Pro is multimodal — you can drag an image into your terminal.
Ad esempio:
Here is the sketch for “Project Constellation.” Generate the full UI in HTML/CSS/JS.
@sketch.png
The CLI will detect interface elements (buttons, cards, layout) and convert your drawing into code.
5. Improve Daily Developer Workflows
Generate Shell Commands
You can describe an action in natural language:
Find the commit that changed my default theme to dark using git bisect.
Gemini CLI will:
- create the bisect sequence
- run each step
- return the commit hash
Generate Documentation From Code
Review all the code and write complete user-facing documentation with
feature explanations, CLI options, architecture overview, and contribution guide.
Gemini 3 Pro will analyze the codebase and create a structured, readable document.
Debug Cloud Run Performance Issues
Gemini CLI can orchestrate multi-step workflows:
Users report that the “Save Changes” button is slow. Investigate the 'tech-stack'
Cloud Run service and identify the root cause.
Gemini will:
- gather logs
- check resource usage
- analyze security scans
- propose a fix
- deploy the fix if confirmed
Conclusione
Gemini 3 Pro transforms the terminal into a powerful AI development environment. Whether you’re building full-stack apps, debugging infrastructure, or translating sketches into functional UI, Gemini CLI + Gemini 3 Pro delivers state-of-the-art reasoning and productivity.
