GPTPRO

How to Use Gemini 3.5 Flash? 5 Ways to Quickly Experience Google's Newest I/O Model + Tutorial

🚀 Google I/O just dropped a brand-new generation of LLMs, and everyone is searching: how to use Gemini 3.5 Flash? To help developers and AI enthusiasts get hands-on with this model — built for "blazing speed and high cost-performance" — as quickly as possible, this article rounds up the latest and most complete 5 ways to use Gemini 3.5 Flash.

Whether you're a regular user looking for a free ride, a beginner hunting for a no-VPN direct-access solution in China, or a hardcore geek developer, this step-by-step guide doesn't just teach you how to use Gemini 3.5 Flash — it also includes a detailed "Heavyweight Pitfall Guide" to help you clear every hurdle and experience the full power of Gemini 3.5 Flash with zero friction!


📑 Table of Contents (Click to Jump)


Method 1: Upgrade Antigravity IDE to 2.0

69

For developers wondering how to use Gemini 3.5 Flash for coding, the quickest route is through the Antigravity IDE.

If you're already using the Antigravity environment, just open the client, upgrade to the latest 2.0 version, then open the model picker at the top of the editor and switch to Gemini 3.5 Flash to start.

⚠️ Upgrade Pitfall Guide:

  • 🚨 Pitfall 1: Google account login fails / can't verify Some users have found they can't sign into their Google account after upgrading. This seems to be caused by cache issues or fluctuating regional IPs. Solution: Try fully signing out and back in, or switch to a cleaner global proxy environment.

  • 🚨 Pitfall 2: The IDE "disappeared" after upgrading the main client Antigravity 2.0 is a major update that split the IDE out of the main client into a standalone App. If you can't find the coding workspace after upgrading the main client, it's because the new standalone app isn't installed locally — you'll need to download and install it separately. 70

  • Official download link: https://antigravity.google/download#antigravity-ide


Method 2: Use Antigravity CLI

If you prefer hammering on the keyboard in a terminal, the CLI command line is an excellent way to experience Gemini 3.5 Flash. Essentially, the old Gemini CLI has been upgraded to Antigravity CLI — once installed, just run the agy command to start interacting.

⚠️ Prerequisite note: Before using the CLI, you must first complete login authorization with an Antigravity account or Antigravity IDE; otherwise the CLI can't run model conversations.

💻 macOS / Linux installation command: Run the following directly in the terminal:

bash
curl -fsSL https://antigravity.google/cli/install.sh | bash

💻 Windows PowerShell installation command:

71

powershell
irm https://antigravity.google/cli/install.ps1 | iex

73

After installation, close the current terminal and reopen it. Type agy --version to check whether the install succeeded, then type agy to sign in. 74

In the prompt that appears in the terminal, choose option 1 (Google OAuth login). 75

The browser will then automatically open the Google login authorization page (if it doesn't, copy the link generated in the terminal and open it manually in the browser to complete authorization).

💻 Windows CMD installation command:

cmd
curl -fsSL https://antigravity.google/cli/install.cmd -o install.cmd && install.cmd && del install.cmd

Once installed, type agy in the terminal to launch it. If all goes well, you'll see the Gemini 3.5 Flash command-line interface.


Method 3: Google AI Studio

77

Many netizens search for "how to use Gemini 3.5 Flash for free" — and Google AI Studio is definitely your first choice! It's currently the most non-paying-user-friendly, highest success-rate direct-access channel from China, and the most stable official path.

  1. Open the site and sign in with your regular Google account.
  2. In the "Model" picker on the right panel, switch to Gemini 3.5 Flash (Preview) (note: if some interfaces show the Gemini 3 Pro preview, just select the Flash version as needed).
  3. Start right away! The platform supports an ultra-long context of up to 1 million tokens, lets you upload long videos and multi-page PDF papers, and even drag in an entire GitHub repo for the AI to analyze.

💡 Key advantage: No extra paid subscription needed to use the latest model, no strict API rate limits, and the strongest native multimodal (vision/audio/text) capability — the best free platform for experiencing Google I/O's newest AI tech.


Method 4: Gemini Official Web & App

76

For users accustomed to the native ChatGPT-style web interface or the mobile native App, you can also use Gemini through official consumer channels directly, though there are some entry thresholds to be aware of.

  • 🌐 Web URL: https://gemini.google.com
  • 💳 Entry requirements: Free accounts can only use the basic model by default. You'll need to subscribe to Google AI Pro (about $20/month) or a higher Ultra plan to unlock the latest Gemini 3.5 Flash / Pro advanced model family on the official web portal.
  • ✅ How to use: If you're an existing Plus/Pro subscriber, simply open the model dropdown at the top of the web or App and select the newly added Gemini 3.5 Flash model to start fast conversations.

Method 5: China Mirrors / Aggregator Platforms (No-VPN Direct Access)

If you're facing network restrictions and don't want to wrestle with complex network environments and proxy tools, the best answer to "how to use Gemini 3.5 Flash in China" is to use a high-quality third-party API aggregator platform — these platforms have already plugged in Gemini 3.5 Flash (preview API) at lightning speed.

  • Recommended test site: https://chatshare.biz
  • Platform advantages: Accessible directly from China's regular network environment; supports one-click seamless switching between Gemini 3.5 Flash, GPT-5.5, Claude 4.7 and other world-class mainstream models. Great for light users and researchers who need multi-model comparisons.

❓ FAQ & Heavyweight Pitfall Guide

To help everyone avoid detours while exploring "how to use Gemini 3.5 Flash," here's a troubleshooting guide for recent hot-button issues:

Q1: What's the difference between Antigravity IDE and the Antigravity main client?

A: Since the big 2.0 update, they've become completely separate apps. The IDE is dedicated to writing code, debugging, and integrating AI-assisted programming environments; the main client handles Google ecosystem account management and other non-coding ecosystem features. For the full AI-coding experience, we recommend installing both apps.

Q2: Why does typing agy in the terminal after installing the CLI show "Command not found"?

A: This usually happens because the install script failed to write the executable's environment variable path (Path) into your shell configuration. ✅ Solution: Manually refresh the environment variable config. Run source ~/.bashrc or source ~/.zshrc in the terminal (depending on which shell your Mac/Linux uses), then type agy again and it will work.

【🚨 Ultimate Pitfall】Q3: After upgrading to 2.0, all my plugin configs and chat history are gone. What do I do?!

A: Don't panic! This absolutely isn't data loss — it's a local file storage path change caused by 2.0 splitting the IDE into a standalone app. Simply rename two hidden folders manually and all your original config and data will be recovered intact:

For macOS / Linux users, run or manually complete these two steps in the terminal:

  1. Rename the old directory /Users/YourUsername/Library/Application Support/Antigravity to Antigravity IDE
  2. Rename the old directory /Users/YourUsername/.antigravity to .antigravity-ide

✅ Recovery result: Once the folders are renamed, restart your Antigravity IDE and you'll be pleasantly surprised to find that your previously installed third-party plugins and precious AI chat history are all back, exactly as they were!

Share this article

Related reading