Google rolls out Antigravity Orchestrator: dynamic in-task model switching, 24% less quota
rseroter · x · 2026-11-18
Google is rolling out Orchestrator in Antigravity for AI Ultra subscribers. Unlike a router that picks one model per task, it dynamically switches models within a single task: Carbon sets the plan, hands code edits and testing to the fast Hydrogen, and steps back in to repair the plan on friction — near-Carbon success rates on Terminal-Bench 4.0 while using 24% less quota and finishing 19% faster. Commenters highlight the Orchestrator + /swarm combo.
More from coding & agent
- Scientific computing lead shares multi-model coding workflow: free models write, Claude coordinates — ChrisRackauckas · 2026-11-18
- Capability Is Not Authority: The Hidden Risks of Deploying AI Agents — Adjust-the-rhythm · 2026-11-18
- SciML author shares multi-model coding workflow: role-based routing for writing and review — ChrisRackauckas · 2026-11-18
- AI Agents That Need Constant Babysitting Aren't Agents — Few-Garlic2725 · 2026-11-18
- Dev wants a single client to manage all agent threads with Slack-style swipe UI — curious_vii · 2026-11-18
- Lindy rebuilt its AI teammate engine with coding agents — faster at half the cost, 9 lessons — mattbeane · 2026-11-18