Can Google Antigravity 2.0 make developers pay more for agent-heavy coding before they know how far the new limits actually stretch?
Google used Google IO 2026 to launch Antigravity 2.0, a major update to its agentic coding app with a revamped desktop app, a new terminal-based CLI tool, and an SDK for custom workflows, according to TechCrunch. The release lands alongside a new $100 AI Ultra plan that gives users 5x higher AI limits in Antigravity than the AI Pro plan.
That pairing is the real story. Google is not just shipping another AI coding interface. It is tying developer agents, background automation, and heavier usage to a clearer paid ladder.
Can Antigravity 2.0 turn Google’s coding agent into a daily developer workspace?
Antigravity 2.0 is Google’s answer to the developer who wants AI agents to do more than autocomplete code. The updated desktop app lets users orchestrate multiple agents, run tasks simultaneously, design custom subagent workflows, and schedule tasks that run automatically in the background, TechCrunch reported.
That moves Antigravity closer to an operating layer for AI-assisted software work. Instead of one prompt, one answer, one edit, Google is pushing a model where multiple agents can work across jobs while the developer supervises.
The new app also connects more directly with Google AI Studio, Android, and Firebase. That matters because Google’s advantage is not only the model. It is the stack around the model.
A developer building an Android app, testing Firebase integration, and experimenting in AI Studio is exactly the kind of user Google wants to keep inside its own tooling.
Google said much of the new Antigravity experience is powered by Gemini 3.5 Flash, which it said was co-developed using Antigravity. That creates a feedback loop: the coding tool helps develop the model, and the model powers the coding tool. We previously covered the broader strategic bet behind that model in Cheap AI Agents: Google’s Gemini 3.5 Flash Bets Big.
The new Antigravity CLI is aimed at programmers who prefer terminal workflows. Google is asking users of the Gemini CLI to migrate to the new tool, which signals that Antigravity is becoming the center of its agentic coding push rather than a side project.
Google is also adding native voice command support to Antigravity, mirroring voice features it has brought to consumer products including Gmail and Docs, alongside efforts like Google Pics AI design. For developers, voice may be less important than the CLI. But it shows Google wants Antigravity to span multiple input styles, not just chat boxes.
Does the $100 AI Ultra tier solve the real bottleneck: usage limits?
The new AI Ultra plan costs $100 and offers 5x higher AI limits in Antigravity than the Pro plan.
Google says the new $100 AI Ultra plan offers 5x higher AI limits in Antigravity than the Pro plan.
That is a direct pitch to power users. Agentic coding can burn through usage faster than a conventional coding assistant because the tool is not simply answering questions. It may be running parallel agents, generating code, testing paths, scheduling work, and building subagent workflows.
Google is also cutting the price of its top AI Ultra plan from $250 to $200. That higher tier allows 20x higher limits than the Pro plan, according to TechCrunch.
| Google AI plan detail | Confirmed change |
|---|---|
| New AI Ultra plan | $100 price point |
| $100 AI Ultra limit | 5x higher AI limits in Antigravity than Pro |
| Top AI Ultra plan | Reduced from $250 to $200 |
| $200 AI Ultra limit | 20x higher limits than Pro |
The pricing structure matters because Antigravity 2.0 encourages more frequent AI-assisted sessions through both desktop and terminal access. A developer who only uses AI for occasional code explanations may not need the new ceiling. A developer running multiple agents and background tasks could hit limits much faster.
This is also where Google is aligning with the paid AI tooling market described by TechCrunch. Other AI labs, including Anthropic and OpenAI, have added $100- and $200-per-month plans in recent years for users with different AI usage requirements.
MLXIO analysis: Google’s move is less about undercutting rivals and more about normalizing higher monthly spend for AI-native development workflows. The product asks developers to use more automation. The pricing then charges for the headroom that automation requires.
Can desktop plus terminal access make Antigravity harder to ignore?
The updated desktop app and CLI give Google two routes into developer habits.
The desktop app targets visual orchestration: multiple agents, background tasks, project integrations, and custom workflows. The CLI targets the place many programmers already live: the terminal.
That split is important. If Antigravity only worked as a polished desktop tool, it could miss developers who want fast, scriptable, local-feeling workflows. If it only shipped as a CLI, it could miss teams that need visibility into what agents are doing across tasks.
Google is also launching an Antigravity SDK for developers who want to build custom agents based on its coding tool. Google Cloud customers will be able to connect to Antigravity to build projects, and Google said it will release custom agent templates in AI Studio for enterprise users.
The company is adding an Antigravity export tool to AI Studio so developers can export an existing project and continue working locally. That is a practical bridge between experimentation and actual development work.
The bigger pattern reaches beyond coding. Google is using Antigravity’s coding capabilities in Search, where users will get a custom UI generated in real time as part of an answer and will be able to build mini-apps while exploring a topic, the company said. That connects with our earlier coverage of AI Agents Grab Google Search—and Start Watching You, where Google’s agent strategy began moving deeper into the search experience.
Which Antigravity details will decide whether developers actually switch?
The announcement leaves several practical questions unanswered.
TechCrunch’s report does not specify a detailed rollout schedule, supported desktop platforms, or the full technical limits of the new CLI and SDK. It also does not spell out how the new Antigravity CLI will map onto existing Gemini CLI workflows beyond Google asking Gemini CLI users to migrate.
For developers, the real test will be less theatrical than the launch. It will come down to:
- Reliability: Whether multiple agents can run without creating messy handoffs or unusable code.
- Latency: Whether terminal and desktop workflows feel fast enough for daily use.
- Accuracy: Whether Gemini 3.5 Flash can handle coding tasks without constant correction.
- Limits: Whether the 5x and 20x allowances reduce interruptions for heavy Antigravity users.
- Migration: Whether Gemini CLI users can move without breaking existing habits.
The updated desktop app gives Google a broader surface area. The CLI gives it developer credibility. The SDK gives teams a reason to customize. The pricing gives Google a way to monetize the heavier usage those features are designed to create.
The next watch item is adoption, not announcement volume. If Antigravity 2.0 turns parallel agents, scheduled coding tasks, and local project export into reliable daily workflows, the $100 AI Ultra tier could become a practical developer expense. If the limits remain opaque or the CLI migration feels rough, the launch risks looking like another ambitious AI tool that asks for commitment before proving it can handle production pressure.
The Bottom Line
- Google is turning Antigravity from a coding assistant into a broader agentic developer workspace.
- The new $100 AI Ultra plan ties heavier coding-agent usage to a clearer paid upgrade path.
- Deeper links with AI Studio, Android, and Firebase could keep developers inside Google’s ecosystem.
