At Google I/O 2026, Google unveiled Gemini 3.5 Flash, the first model in the Gemini 3.5 family. The company says it delivers four-times-faster output than competing frontier models, improves coding and agentic performance beyond Gemini 3.1 Pro, and is rolling
By the time Google’s Gemini demo started moving fast, the pitch was clear: this isn’t just another model release—it’s speed meant to feel practical.
At Google I/O 2026, Google introduced the Gemini 3.5 family of AI models, beginning with Gemini 3.5 Flash. Google said the model is four times faster than competing frontier models when measured by output tokens per second. while also outperforming Gemini 3.1 Pro on coding and agentic benchmarks. The message from the stage was that “Flash” isn’t chasing novelty; it’s trying to make coding and long-horizon. agent-style work feel usable at the pace people expect.
Google says Gemini 3.5 Flash combines what it calls frontier intelligence with action—so the model can not only perform advanced reasoning and deep logical thinking. but also move toward autonomously executing long-horizon tasks. The company positions that balance of speed and capability as the reason it’s well-suited for work such as developing new applications. maintaining codebases. and preparing financial documents.
Beyond raw performance, Google also claimed Gemini 3.5 Flash improves the look and feel of outputs. The company said it can generate richer, more interactive web UIs and graphics. It also said the broader Gemini 3.5 effort is less likely to generate harmful content or mistakenly refuse to answer safe queries. pointing to new safety training and mitigations. along with interpretability tools designed to help check and understand the AI’s inner reasoning before it responds.
What matters for everyday users is access. Gemini 3.5 Flash is already rolling out to billions of users globally as the default model in the Gemini app and in AI Mode in Google Search.
For people building with it. Google’s also made the model available through several developer and enterprise paths: Google Antigravity (Google’s agentic development platform). the Gemini API for Google AI Studio and Android Studio. and enterprise access via the Gemini Enterprise Agent Platform and Gemini Enterprise.
Google is also tying Gemini 3.5 Flash into its personal-agent push. The company said it’s using Gemini 3.5 Flash in Gemini Spark. a personal AI agent that runs 24/7 to help navigate a user’s digital life and take actions on the user’s behalf under their direction. Gemini Spark is rolling out to trusted testers today. with a wider beta planned for Google AI Ultra subscribers in the US next week.
The release doesn’t stop at Flash. Google said it is already working on Gemini 3.5 Pro, with employees using the Pro model internally, and a goal to roll it out next month.
Taken together. the rollout plan makes one bet: if speed is the bottleneck for coding and agentic workflows. then making Flash both quicker and stronger—while expanding access across consumer. developer. and enterprise tools—could change how often people actually rely on these systems to do real work.
Gemini 3.5 Flash Google I/O 2026 Google Gemini AI models coding benchmarks agentic benchmarks Gemini app Google Search Gemini API Google Antigravity Android Studio Gemini Enterprise Gemini Spark AI Mode AI Ultra