Google introduces Antigravity 2.0, featuring an updated desktop app that lets users orchestrate agents, an Antigravity CLI tool, and an SDK for custom workflows
Google is introducing a new version of its agentic coding app, Google Antigravity 2.0, with an updated desktop app, a CLI tool, and an SDK for custom workflows.
TechCrunchIvan Mehta
Context & Ripple Effects
Antigravity began as a public-preview, agent-first coding tool built around Gemini 3 Pro and third-party models. Google was also reported to be consolidating its internal AI coding tools under the Antigravity platform in response to Claude Code and Codex.
The 2.0 release broadens that product from an individual coding tool into a workflow surface: a desktop interface for coordinating agents, a command-line interface, and an SDK. It aligns with Google’s parallel effort to manage agent fleets through its Gemini Enterprise Agent Platform.
First-order effects
Developers using Antigravity gain more ways to run agentic coding work: desktop orchestration for multiple agents, terminal-based access through the CLI, and an SDK for tailored workflows.
Google gets a more complete Antigravity distribution and development surface, rather than relying on a single agent-first coding application.
Second-order effects
Teams can integrate Antigravity into existing developer workflows and build custom automations, increasing the practical importance of interoperability between Google’s coding agents and other tools.
The broader product surface raises pressure on competing coding-agent offerings such as Claude Code and Codex to match not only model capabilities but also orchestration, command-line, and extensibility options.
Third-order effects
If Google continues connecting Antigravity to products such as NotebookLM and its enterprise agent platform, coding assistants may evolve into a shared agent runtime across developer and knowledge-work products rather than remain standalone chat or IDE features.
The competitive boundary could shift from the quality of a single coding model to control of the workflow layer—agent coordination, custom integrations, and lifecycle management—though adoption will determine whether that platform strategy takes hold.
The trend: Antigravity 2.0 is part of the shift from isolated AI coding assistants toward extensible platforms for coordinating fleets of agents across developer workflows.
Just off stage at #GoogleIO, some highlights from this morning 🧵 Gemini 3.5 Flash is available today for everyone in @antigravity and across our products and APIs. Compared to 3.1 Pro, 3.5 Flash is better across almost all benchmarks with huge progress in coding. It's also [image…
We asked our agents to build a working operating system from scratch using @Antigravity 2.0 and Gemini 3.5 Flash. It took: ⏱️ 12 hours 🤖 93 parallel sub-agents 🔄 15k+ model requests 🧠 2.6B tokens processed 💸 Less than $1K in API credits To build a functioning OS from [video]
Google built an entire operating system using Antigravity and Gemini 3.5 Flash for under $1,000 in API credits It took 93 subagents, 12 hours, 15K model requests, 2.6B tokens...
All agentic coding apps are starting to look the same. This is the latest Antigravity app from IO, which looks pretty slick. But I think this UI is only optimized for a single person to work to agents. The UI for teams and orgs to work well with agents still needs to be [image]
Google released Antigravity CLI as the only way to use CLI now (3.5 Flash is not available in Gemini CLI anymore). And ofcourse, they do not migrate settings/skills/MCPs anything from Gemini CLI to Antigravity CLI automatically. Classic.
We're expanding the @Antigravity ecosystem to help developers shift their energy away from tedious debugging and focus on what actually matters: architecture and design. 🧵 [image]
Google Antigravity is expanding, including a new standalone desktop app that acts as a central home for agent interaction. We're also introducing a new Antigravity CLI providing a fast, lightweight way to deploy new agents instantly without a graphical user interface, and a new […
The agent-first @antigravity takes a huge step forward. Standalone desktop app, CLI, SDK, and a harness across surfaces. Dynamic sub agents! We built an OS from scratch for $1000 in tokens. Quite a test! #googleio [image]
Google must be running the most insane data centers on earth with the new Antigravity (nice work to the team, it's night and day) — and Gemini 3.5. It's so good, and so fricking fast, I have to wonder why they didn't call it Gemini 4.0. It's simply amazing..
Clearly Antigravity is not SOTA, still using /grill-me and not /grill-with-docs Side note - pretty wild to see a skill I cooked up being used in a Google product demo
Google Antigravity 2.0 is interesting because it no longer feels like “Google made an AI IDE”. Antigravity 1.0 was the full IDE: editor, terminal, browser, agent workspace. Basically Google's take on agentic coding as a complete environment. 2.0 feels more like they pulled the [i…
Deleted again because misinformation 🥲 Gemini 3.5 Flash *is* available on the API. Yet, both the API and the CLI versions are 3x slower than on the IDE! See the video below. → Antigravity IDE: 4 seconds (smooth) → Antigravity CLI: 15 seconds (buggy) So the point holds: [video]
I don't have high hopes for Antigravity - or any dev tool Google ships, outside of possibly ones in Chrome - because Google seems to have no DNA in dev tools. They start a project that has promise, and a year or two later half the team will do an internal transfer, to work on
I had to do a double-take: in the second minute of the launch video for Antigravity 2.0 you can see people use Codex, on the Antigravity team Did no one double check the launch video, the very least? This is typical Google btw... I don't have high hopes for Antigravity either [im…
Google coding agent devs accidentally leaking to the world that they are using OpenAI Codex in a marketing video is hilarious. Google's stifling bureaucracy is legendary. I still proclaim Google's agentic coding efforts DOA [image]