Google’s new Gemini 3.5 Flash promises better coding and agents

Gemini is one of the best AI models around, with each release being widely applauded for pushing the envelope of what end users can achieve with AI. Google released Gemini 3 in November 2025, followed by Gemini 3.1 in February 2026. Today at Google I/O 2026, Google is releasing Gemini 3.5 to the world, starting with Gemini 3.5 Flash, which is already rolling out to billions of users globally.

Google says the latest Gemini 3.5 family of AI models combines frontier intelligence with action. This means users can expect these models to not only execute advanced reasoning and deep logical thinking, but also move towards autonomously executing long-horizon tasks.

Gemini 3.5 Flash: Features and highlights

Google claims that Gemini 3.5 Flash delivers intelligence that rivals large flagship models on multiple dimensions, at the speeds users have come to expect from the Flash series. It’s also the company’s strongest agentic and coding model yet, outperforming even Gemini 3.1 Pro on challenging coding and agentic benchmarks. When looking at output tokens per second, it is four times faster than other frontier models.

The balance of speed and performance makes Gemini 3.5 Flash ideal for tackling long-horizon agentic tasks. It can rapidly plan, build, and iterate to solve real-world problems, whether it’s developing new applications, maintaining codebases, or preparing financial documents.

Gemini 3.5 Flash also generates richer, more interactive web UIs and graphics.

Google says that Gemini 3.5 as a whole is less likely to generate harmful content and mistakenly refuse to answer safe queries. Google achieves this with new, more advanced safety training and mitigations, including interpretability tools that help check and understand the AI’s inner reasoning before it responds.

Gemini 3.5 Flash is available today to billions of people globally through the Gemini app and AI Mode in Google Search, where it is the default model.

For developers, Gemini 3.5 Flash is available in Google Antigravity (Google’s agentic development platform) and in the Gemini API for Google AI Studio and Android Studio. It is also available to enterprise users via the Gemini Enterprise Agent Platform and Gemini Enterprise.

Google is also using Gemini 3.5 Flash in Gemini Spark, a personal AI agent that runs 24/7 to help you navigate your digital life, taking actions on your behalf under your direction. Gemini Spark is rolling out to trusted testers today, with a wider beta arriving for Google AI Ultra subscribers in the US next week.

Google also mentioned that it is already hard at work on Gemini 3.5 Pro. Employees are using the Pro model internally, and Google hopes to roll it out next month.