Source: Google BlogMay 19, 2026

Google I/O 2026: Gemini 3.5 Flash Delivers Frontier Intelligence at Flash Speed

View original source →

Google announced Gemini 3.5 Flash at I/O 2026 on May 19, positioning it as the first model in a new series that combines frontier-class intelligence with the speed and cost profile of the Flash lineup.

The model closes the traditional trade-off between intelligence and latency, making it the first frontier-class model viable for real-time agentic applications.

Key points:

• Gemini 3.5 Flash outperforms Gemini 3.1 Pro across coding, agentic task execution, and multimodal reasoning benchmarks

• Output delivery at 4x the tokens-per-second of other frontier models at comparable intelligence levels

• Gemini 3.5 Pro is in testing and will be available next month; Flash is available immediately via the Gemini API

• Flash-tier pricing democratizes access to high-capability AI for startups and developers previously priced out of frontier deployments

The intelligence-at-speed combination directly challenges OpenAI's GPT-5 series and Anthropic's Claude Sonnet line for enterprise agentic workloads where latency is a practical adoption barrier.

Developers building agentic applications should evaluate Gemini 3.5 Flash as a primary model immediately. The latency-intelligence combination is purpose-built for multi-step workflows where speed compounds across tool calls.

Why It Matters: A frontier-class model at Flash pricing expands Gemini's addressable market immediately. For enterprise agentic workloads, eliminating the latency penalty changes which applications are practically viable.

Google I/O 2026: Gemini 3.5 Flash Delivers Frontier Intelligence at Flash Speed | AI Onboarded