Claude Opus 5.5 and Claude Sonnet 5.5 are now available in Google Antigravity

Google Cloud has integrated Claude Opus 5.5 and Claude Sonnet 5.5 into Antigravity, its agentic development platform, to provide developers with model options tailored to different workloads. Claude Opus 5.5 is positioned for complex tasks requiring deeper reasoning, while Sonnet 5.5 offers a faster, lower-cost alternative for routine or less demanding work. Administrators can enable these models for teams, allowing developers to choose them directly from the model picker alongside Google’s Gemini models in Antigravity 2.0, CLI, and IDE extensions for VS Code, JetBrains IntelliJ, and Visual Studio.

The availability of multiple model families within Antigravity supports multi-model workflows, which can improve code quality and help control latency and costs. For example, using distinct models for implementation and validation can reduce blind spots by providing independent verification before progressing in a workflow. This approach is particularly useful in agentic loops or multi-step graph workflows where different stages require different levels of reasoning.

Antigravity’s Subagent customization system allows teams to assign specific model tiers to specialized subagents based on task requirements. High-speed search, classification, and triage work can be offloaded to faster models like Gemini Flash, while deeper reasoning tasks remain with more powerful models such as Opus 5.5. This setup enables efficient delegation of subtasks while optimizing both performance and cost.

The integration also maintains lightweight first-party Gemini models in the background for high-frequency client-side tasks like context management and summaries. These models operate independently and may incur separate charges, as they consume first-party tokens. Developers can refer to the full list of supported models in Antigravity documentation for subscription-specific details.