HaiAI123

Curated Global AI Tools Directory

Industry update
Flat

Google Opens Antigravity /teamwork Multi-Agent Tool to Researchers and Developers

11/14/2026 — 11/16, 05:16·3 sources (all)·3 reports (all)

Key takeaways

  • Google Research has opened its Antigravity /teamwork multi-agent system to researchers and developers via the /teamwork interface. Built on…

Background & analysis

On November 14, 2026, Vahab Mirrokni, a Google Fellow and VP at Google Research, announced that the Antigravity /teamwork multi-agent system, built on Gemini 4 Argon, Gemini 4 Carbon, and Gemini 4 Hydrogen, had been working for several weeks with internal Google teams and academic partner institutions to tackle long-open problems in mathematics, chemistry, and complex systems engineering. Google Research said these agent tools were being made available externally through /teamwork, so researchers and developers could apply them to their own projects.

The same day, IT之家 reported that Google was expanding the Antigravity multi-agent system with Teamwork, a suite of specialized agents driven by Gemini models for verifying complex mathematical proofs and optimizing software performance. According to the report, the Teamwork-LEAP system completed a formalization of the Weil conjectures through roughly 28,800 sub-agent calls and about 1.8 million lines of Lean code. Google also said the system reproduced a formal proof concerning finite-time blow-up for the three-dimensional Navier-Stokes equations, and helped researchers resolve the Courtade-Kumar conjecture. Architecturally, the system runs multiple agents in parallel and assigns an independent system to challenge and verify proposed solutions, limiting errors that a single AI model might produce.

On November 16, DEV Community · AI relayed Google Research's claim that teams of Gemini 4 agents inside Antigravity proved the Courtade-Kumar conjecture, originally posed in 2013, rewrote a video decoder in Rust for a 20-30% speedup, and ported a microkernel to safe Rust. The write-up described /teamwork as Antigravity's multi-agent mode, in which producers, critics, and auditors divide work to solve hard problems. The materials do not detail how access requests are handled or how the open tools are currently being adopted.

AI-generated from 3 reports · updated 1 day ago

Latest turnConvertElementor relays a Google Research claim that teams of Gemini 4 agents inside Antigravity proved the 2013 Courtade–Kumar conjecture, rewrote a video decoder into Rust for a 20–30% speedup, and ported a microkernel to safe Rust. /teamwork is Antigravity's multi-agent mode, splitting a hard problem across producers, critics and auditors.

Related AI tools

updream
24-hour heatHeat index 40 · peak 105 · 23h ago
24 hours agonow

Reports on this story headlines open the original

Yesterday
  1. ConvertElementor relays a Google Research claim that teams of Gemini 4 agents inside Antigravity proved the 2013 Courtade–Kumar conjecture, rewrote a video decoder into Rust for a 20–30% speedup, and ported a microkernel to safe Rust. /teamwork is Antigravity's multi-agent mode, splitting a hard problem across producers, critics and auditors.

    DEV Community · AIAI score 54
Nov 14
  1. IT Home reports that Google has expanded its Antigravity multi-agent system with Teamwork, a setup of specialised Gemini-powered agents that verify complex math proofs and optimise software performance. Google says the system used about 28,800 sub-agent calls and roughly 1.8 million lines of Lean code to formalize the Weil conjectures, and also reproduced a formal proof on finite-time blowup for the 3D Navier-Stokes equations.

    IT之家AI score 64
  2. Google Research says its Antigravity /teamwork multi-agent system, built on Gemini 4 Argon, Carbon and Hydrogen, helped internal and academic teams make progress on long-open problems in information theory, formal verification, molecular structure discovery and kernel optimization. The agentic tools are now available externally through /teamwork so researchers and developers can apply them to their own work.

    Google ResearchFirst-partyAI score 62

Other stories people are talking about

How is heat calculated?About the method

Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.

This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.

Surge
Discussion rising fast
New
First report within 6 hours
Rising
Still gathering discussion

Back to AI industry updates →