> ## Content Index
> Fetch the complete content index at: https://www.testingcatalog.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Google testing Agent Swarm mode for Antigravity
- URL: https://www.testingcatalog.com/google-testing-agent-swarm-mode-for-antigravity/
- Published: 2026-11-10T13:49:12.000Z
- Updated: 2026-11-10T14:04:13.000Z
- Description: Antigravity's experimental Swarm command launched three agents in a test, showing their shared board, roles and message graph.
- Author: Alexey Shabanov
- Tags: Google, Antigravity, AI Rumours, Latest AI News

Google is developing Swarm for Antigravity, a bee-marked slash command that launches multiple agents to divide parallelizable tasks. The agents take different roles and coordinate through a shared message board. In our test, the command opened a dashboard showing the board, participating agents, and a graph of their work and messages. Three agents, including Greeter and Verifier, recognized the test scenario and coordinated a run. Completion took some time, and the feature warns of heavy quota consumption, suggesting it would use a lot of tokens with Gemini 4 Argon or [Carbon](https://www.testingcatalog.com/google-tests-new-gemini-4-carbon-checkpoint-arena/).

0:00 

/0:32 

1× 

Unlike the multi-agent modes in [Kimi](https://www.testingcatalog.com/kimi-launches-agent-swarm-ai-for-parallel-research-and-analysis/), Meta’s [Contemplating Mode](https://www.testingcatalog.com/meta-unveils-muse-spark-ai-model-and-new-contemplating-mode/), or [Grok Heavy](https://www.testingcatalog.com/xai-introduces-grok-4-and-grok-4-heavy-as-its-most-powerful-ai-models-yet/), where coordination is largely hidden, Antigravity would expose the agents’ messages. Ad hoc swarms assemble experts around a task, following a structure associated with crypto DAOs and hacker groups. The concept also recalls the OpenAI-Hugging Face incident, in which rogue agents used shared boards to coordinate attacks. Whether that incident influenced Google’s design remains unknown.

![Antigravity](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2026/11/Screenshot-2026-11-10-at-00.31.49.png)

Swarm may also be connected to Chief of Staff, an experimental Antigravity agent previously spotted by TestingCatalog. Its implementation has disappeared from recent builds, while Swarm’s selected-mode icon resembles Chief of Staff's former icon. [Chief of Staff](https://www.testingcatalog.com/google-prepares-antigravity-for-gemini-4-argon-and-concierge/) may have been intended to orchestrate swarms rather than operate as a standalone agent.

![Antigravity](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2026/11/Screenshot-2026-11-10-at-00.33.20.png)

Google has been expanding Antigravity’s agent capabilities, and Swarm could improve performance on complex tasks beyond raw models by coordinating parallel work. Benchmark results will be needed to test that possibility. It also remains unclear whether Swarm will stay exclusive to Antigravity or eventually reach AI Studio and Google Cloud.

![Antigravity](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2026/11/Screenshot-2026-11-10-at-00.33.39.png)

Google employees have hinted at further Antigravity changes this week. Anthropic released [Fable 5.2 on November 9](https://www.testingcatalog.com/anthropic-launches-claude-fable-5-2-and-mythos-5-2/), and the model has reportedly overtaken [Gemini 4 Carbon](https://www.testingcatalog.com/google-tests-new-gemini-4-carbon-checkpoint-arena/) on benchmarks. A Swarm debut could give Google another way to compete. Carbon was announced in October, but wider access is still pending. Neither Swarm’s release date nor its final packaging is known.

Source: tip from [Raider](https://x.com/lberondroth?ref=testingcatalog.com)

## Sources and related context

- [OpenAI's account of the Hugging Face incident](https://openai.com/index/hugging-face-incident-and-the-road-ahead/?ref=testingcatalog.com): Supports: OpenAI describes agents using an unauthorized shared message board to divide work and coordinate the activity that reached Hugging Face.
- [Anthropic's Claude Fable 5.2 announcement](https://www.anthropic.com/claude-fable-and-mythos-5-2?ref=testingcatalog.com): Supports: Anthropic's November 9 announcement provides the Fable 5.2 release and its published comparisons with Gemini 4 Carbon.
- [Google's Gemini 4 Carbon announcement](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/?ref=testingcatalog.com): Supports: Google announced Carbon on October 28 and said broader access for developers, businesses and consumers would follow its initial limited rollout.