Google Antigravity

Google Antigravity is Google's agent-first development platform, launched at Google I/O 2026 as the next evolution of AI-assisted coding. Built as a modified fork of VS Code, Antigravity moves beyond code completion and chat assistance to deploy autonomous agents that plan, execute, and verify complex development tasks across editor, terminal, and browser in parallel. The platform is available as a desktop application for macOS, Windows, and Linux, a command-line interface, an SDK for custom workflows, and Managed Agents in the Gemini API. Antigravity introduces a concept called parallel sub-agents: specialized agents for frontend design, backend logic, testing in a live Chrome instance, and deployment work simultaneously on different aspects of a task. Rather than reviewing raw tool call logs, developers verify agent work through Artifacts: tangible deliverables including task lists, implementation plans, screenshots, and browser recordings that provide the necessary context to validate the agent's logic at a glance. Developers can leave feedback directly on an Artifact, similar to commenting on a document, and the agent incorporates the input without stopping its execution flow. The platform supports multiple model families. The free Individual tier includes Gemini 3.5 Flash, Gemini 3.1 Pro, Gemini 3 Flash, Claude Sonnet 4.6, Claude Opus 4.6, and gpt-oss-120b. Google AI Pro adds more generous rate limits and a flexible AI credit pool. Google AI Ultra provides the highest access to Gemini models with significantly higher rate limits. MCP is supported for connecting external tools and data sources, and the VS Code extension ecosystem is compatible, meaning existing extensions work without modification unless they conflict with the agent layer. Pricing is structured through Google AI subscription tiers. Individual at $0/month is generally available with unlimited tab completions, unlimited command requests, and basic weekly rate limits across all supported models. Google AI Pro is available through Google One AI Premium subscription with more generous rate limits and a flexible credit pool. Google AI Ultra at $100/month provides the highest individual access with 5x Pro usage limits. An Organization plan is available for teams requiring an admin dashboard, seat management, and usage reporting. Pro and Ultra pricing is managed through Google AI subscriptions rather than the Antigravity pricing page directly. Google Antigravity is not the right fit for developers who prefer a non-VS Code editor: JetBrains users should evaluate GitHub Copilot or Kiro for native JetBrains support. Teams requiring a proven track record with independent reviews should note that Antigravity has no G2, Capterra, or Gartner Peer Insights reviews yet despite Google's backing. Organizations needing fully offline or self-hosted AI tooling cannot use Antigravity, which requires internet connectivity for model API access. Developers who want tight integration with project management tools like Jira, Linear, or Slack for task assignment should evaluate Devin, which offers native Slack and Linear integrations for asynchronous task delegation. Current state Q2 2026: Google Antigravity 2.0 launched at Google I/O 2026 with the standalone desktop application, CLI, SDK, and Managed Agents API. Gemini 3.5 Flash powers the agent layer with 4x more output tokens per second than other frontier models. The platform supports Gemini, Claude, and open-weight models from the free tier. Managed Agents in the Gemini API allow spinning up a full agent with a single API call in an isolated Linux sandbox with persistent sessions. Antigravity is available in Gemini Enterprise Agent Platform for Google Cloud customers. No independent reviews exist yet on G2, Capterra, or Gartner Peer Insights.