Google used its I/O 2026 keynote on May 19 to release two things at once: the Gemini 3.5 Flash model and version 2.0 of Antigravity, its agentic development environment.

Gemini 3.5 Flash

Per Google, 3.5 Flash generates output tokens roughly four times faster than competing frontier models and posts strong agentic scores — 76.2% on Terminal-Bench 2.1 and 83.6% on MCP Atlas. It is generally available through the Gemini API in Google AI Studio and Android Studio, inside Antigravity, and — per GitHub’s changelog the same day — as a model option in GitHub Copilot.

The price is the part worth reading twice. Gemini 3.5 Flash lists at $1.50 per million input tokens and $9 per million output, which Google’s own pricing puts at three times the rate of Gemini 3 Flash Preview and six times 3.1 Flash-Lite. “Flash” has long been Google’s shorthand for the cheap, fast tier; 3.5 keeps the speed and drops the cheap.

Antigravity 2.0

Antigravity 2.0 ships as a standalone desktop application built around running several agents at once. Per Google’s announcement it adds dynamic subagents for parallel workflows and scheduled tasks that run in the background, plus Managed Agents in the Gemini API and Android support in Google AI Studio. The pitch is a single place to assign work to multiple agents and let them run.

Where it fits

The two releases push in the same direction: a faster model, and an environment designed to run many agents — scheduled and in parallel. The higher Flash rate is the catch. The per-call cost went up at exactly the moment the tooling is built to make more calls. Whether that nets out as leverage or quiet overspend is taken up in the Flash that got expensive.