Google has launched Antigravity, an agent-first IDE that combines a VS Code fork with multiple AI models and autonomous agents capable of planning, coding, and testing across the editor, terminal, and browser. The platform ships with Gemini 3.1 Pro and Claude Sonnet integrated out of the box, letting developers choose or blend models depending on the task.
The agent architecture is what sets Antigravity apart. Rather than a simple chat-and-autocomplete interface, Antigravity's agents can independently navigate a codebase, open terminals, run commands, interact with browser previews, and iterate on their own output. A developer can describe a feature, and the agent will plan the implementation, write the code across multiple files, run the tests, and fix failures — all without manual intervention.
The pricing model is aggressive: free for individual developers. Google is clearly betting on adoption volume and enterprise upsells rather than individual subscriptions. For the broader AI coding tool market, having Google offer a fully featured agent-first IDE at no cost puts significant pressure on paid tools like Cursor, Windsurf, and others to justify their subscription fees with differentiated features.
