Google is rolling out Gemini 3.5 Flash as the default model behind the Gemini app and AI Mode in Search this week, putting its newest model directly in front of billions of users worldwide.

The switch matters because it changes the model most people will reach when they ask Google a question, and the company says the new Flash is faster and cheaper than the version it replaces. CEO Sundar Pichai told the audience at Google I/O 2026 that “3.5 Flash is better than 3.1 Pro, which was just four months ago, and it’s at almost 90% of the performance of frontier models, 4x faster, and about one-third to one-half the cost.”

That framing, a smaller and cheaper model that closes most of the gap with frontier systems, is the same bet Google made with the original Flash a year ago. The difference this time is scale. Flash now sits at the core of AI Mode in Search, with Personal Intelligence, the feature that pulls in Gmail, Calendar and Drive context, expanding to nearly 200 countries and 98 languages. The company is no longer gating that experience behind a paid subscription.

For developers, the model is generally available through the Gemini API in Google AI Studio and Android Studio, and is plugged into Antigravity, Google’s agent-building platform. GitHub flipped Flash on for Copilot users on May 19, and Google says the model has overtaken its own Gemini 3.1 Pro on coding and agent benchmarks.

The cost story is where Google is making its sharpest pitch to enterprise buyers. Pichai said “a company that uses 1 trillion tokens per day could save more than $1 billion annually by shifting 80% of its workload to Gemini 3.5 Flash.” That math is aimed squarely at customers running AI agents at scale, where token volume, not features, has become the line item that controls the budget.

The trade-off is that smaller, faster models still trail flagship systems on the hardest tasks. Google is positioning Flash as the everyday workhorse, with a heavier Gemini 3.5 Pro queued up for users who need the headroom. In practical terms that means the free Gemini app and AI Mode users get the speed-and-cost optimized version, while paid Gemini Advanced subscribers and enterprise customers will get first access to Pro when it ships.

Personal Intelligence, which the company describes as the connective tissue between Gemini and the rest of your Google account, is the feature most likely to drive day-to-day usage. By pulling in calendar entries, email context and Drive files without a paid tier, Google is betting that habitual Gemini use will follow naturally once the AI can actually see what you’re working on.

A larger Gemini 3.5 Pro is due in June, according to Pichai, who told the I/O crowd to “give us until next month to get it to you.”