Latest News About Gemini 3.5 Flash debuts at I/O

Updated 2026-06-18 06:33

Gemini 3.5 Flash appeared at Google I/O 2026, gaining rapid attention for its speeds near 289 tokens per second and a near million-token context window, enabling long document analysis. It emphasizes agentic tool use and improved multi-step workflow handling while claiming stronger performance in coding, reasoning, and complex task planning than earlier Gemini variants. Reports compare its speed and thinking depth favorably against contemporaries, noting rapid iteration and configurable thinking depth for tasks like math, coding, and large-scale analysis. Early reviews describe it as fast, surprisingly smart, though with questions about practical value in 2026. The model targets coding, multimodal tasks, and agentic workflows, positioning itself as Google’s fastest frontier AI.

Sources