TechBriefly
  • Tech
  • Business
  • Crypto
  • Science
  • Geek
  • How to
  • About
    • About TechBriefly
    • Terms and Conditions
    • Privacy Policy
    • Contact Us
    • Languages
      • 中文 (Chinese)
      • Dansk
      • Deutsch
      • Español
      • English
      • Français
      • Nederlands
      • Italiano
      • 日本语 (Japanese)
      • 한국인 (Korean)
      • Norsk
      • Polski
      • Português
      • Pусский (Russian)
      • Suomalainen
      • Svenska
No Result
View All Result
TechBriefly
Home Tech
Google launches Antigravity for Gemini 3 Pro

Google launches Antigravity for Gemini 3 Pro

Aytun ÇelebibyAytun Çelebi
18 November 2025
in Tech, AI
Reading Time: 2 mins read
Share on FacebookShare on Twitter

Google announced Antigravity, an agent-first coding tool built for Gemini 3 Pro, on November 18, 2025. The tool supports third-party models including Claude Sonnet 4.5 and OpenAI’s GPT-OSS. It provides agents with direct access to code editors, terminals, and browsers. Antigravity enters free public preview for Windows, macOS, and Linux.

Antigravity accompanies the launch of Gemini 3 Pro. Google positions the tool for an “agent-first future.” It supports multiple agents operating simultaneously without constant human intervention. Agents access the code editor to write and edit files, the terminal to execute commands and run scripts, and the browser to interact with web applications and test interfaces.

A central feature is the reporting mechanism. Agents report work plans before starting tasks. As they execute work, Antigravity generates Artifacts. These include task lists outlining objectives, detailed plans specifying steps, screenshots capturing visual states of the workspace or application, and browser recordings documenting interactions within web pages. Artifacts verify completed actions and upcoming intentions. Antigravity also logs actions and external tool usage. Google states that Artifacts are “easier for users to verify” than lists of model actions and tool calls. Developers review these outputs to confirm agent behavior matches expectations.

Antigravity offers two interface views. The Editor view provides an Integrated Development Environment (IDE) experience similar to Cursor and GitHub Copilot. A single agent operates in a side panel alongside the main editing area for real-time assistance during interactive coding sessions. The Manager view enables oversight of multiple agents across various workspaces. Google describes this as “mission control for spawning, orchestrating, and observing multiple agents across multiple workspaces in parallel.” Agents handle tasks independently, with users directing high-level coordination.

Feedback occurs through comments on specific Artifacts. Agents incorporate these comments without pausing ongoing work, preserving workflow continuity while enabling refinements. Agents retain knowledge from previous sessions. They store code snippets for reuse in similar contexts and preserve detailed steps for recurring tasks. This supports progressive improvement as agents build on accumulated experience.

A demonstration video shows Antigravity building a basic flight tracker application. Agents construct the app’s components, conduct tests to ensure functionality, and generate a browser recording of test outcomes. The recording captures the app’s performance in a live browser environment, providing visual proof of execution.

The public preview is free with generous Gemini 3 Pro rate limits that refresh every five hours. Google notes that only “a very small fraction of power users” encounter these limits during typical operation.

Tags: agenticAI agentAntigravityGemini 3Google
ShareTweet
Aytun Çelebi

Aytun Çelebi

Starting with coding on Commodore 64 in elementary school moving to web programming in his teenage years, Aytun has been around technology for over 30 years, and he has been a tech journalist for over 20 years now. He worked in many major Turkish outlets (newspapers, magazines, TV channels and websites) and managed some. Besides journalism, he worked as a copywriter and PR manager (for Lenovo, HP and many international brands ) in agencies. He founded his agency, Linkmedya in 2019 to execute his way of producing content. He is recently interested in AI, automation and MarTech.

Related Posts

Apple begins iPhone 18 series production testing in January

Apple begins iPhone 18 series production testing in January

24 December 2025
EA investigates AI claims in Battlefield 6 cosmetics

EA investigates AI claims in Battlefield 6 cosmetics

24 December 2025
Amazon Alexa+ will book your hotels and salons starting in 2026

Amazon Alexa+ will book your hotels and salons starting in 2026

24 December 2025
OpenAI launches Skills in Codex

OpenAI launches Skills in Codex

24 December 2025

LATEST

How to install mods and custom content in The Sims 2

Running Python files and fixing path errors on Windows

How to boot your PC into Command Prompt for troubleshooting

How to delete a virus using Command Prompt

How to connect a PS4 controller to Steam via USB or Bluetooth

How to connect your phone to Wi-Fi and fix connection issues

Apple begins iPhone 18 series production testing in January

EA investigates AI claims in Battlefield 6 cosmetics

Amazon Alexa+ will book your hotels and salons starting in 2026

OpenAI launches Skills in Codex

TechBriefly

© 2021 TechBriefly is a Linkmedya brand.

  • Tech
  • Business
  • Science
  • Geek
  • How to
  • About
  • Privacy
  • Terms
  • Contact
  • | Network Sites |
  • Digital Report
  • LeaderGamer

Follow Us

No Result
View All Result
  • Tech
  • Business
  • Crypto
  • Science
  • Geek
  • How to
  • About
    • About TechBriefly
    • Terms and Conditions
    • Privacy Policy
    • Contact Us
    • Languages
      • 中文 (Chinese)
      • Dansk
      • Deutsch
      • Español
      • English
      • Français
      • Nederlands
      • Italiano
      • 日本语 (Japanese)
      • 한국인 (Korean)
      • Norsk
      • Polski
      • Português
      • Pусский (Russian)
      • Suomalainen
      • Svenska