Google Introduces Gemini 3.5 Flash: Fastest Frontier Model for Agents and Coding

The Launch

Google DeepMind unveiled Gemini 3.5 at Google I/O 2026 on May 19. The first model in the new family, Gemini 3.5 Flash, is positioned as Google's strongest model for agents and coding, combining frontier intelligence with the ability to take real-world actions.

Technical Specifications

Gemini 3.5 Flash delivers 4x the output speed of comparable frontier models and supports a 1 million-token context window with up to 65,000 tokens of output. Four levels of thinking depth (minimal, low, medium, high) allow developers to trade off cost and performance. The model outperforms Gemini 3.1 Pro on coding and agentic benchmarks.

Availability

Gemini 3.5 Flash is available immediately through the Gemini app and Google's AI Mode in Search. Gemini 3.5 Pro — currently in internal testing — is expected to be available next month.

I/O 2026 Context

The Gemini 3.5 launch was part of a broader set of Google I/O announcements including Gemini Spark (a 24/7 personal AI agent), Gemini Omni (video generation), and Gemini for Science (research tools). Google also cut the AI Ultra subscription from $250 to $200/month and introduced a new $100/month developer tier.