Gemini 3.5 and Gemini Omni: Google's New AI Models from I/O 2026
The headline news from Google I/O 2026 was AI models. Google announced Gemini 3.5, its newest family of models, and Gemini Omni, a new model for creating video. Here’s what was actually released, what’s still coming, and why it matters for your business.
The Gemini 3.5 family
Google describes Gemini 3.5 as “frontier intelligence with action” — models built specifically to carry out multi-step, agentic tasks rather than just answer questions. It’s a family rather than a single model, and Google is releasing it in stages.
One model is available now: Gemini 3.5 Flash. A second, Gemini 3.5 Pro, was confirmed but hasn’t shipped yet — more on that below.
Gemini 3.5 Flash
Gemini 3.5 Flash is the model that launched at I/O. It’s the fast, cost-efficient member of the family, and the notable part is how little you give up for that speed. Google says it delivers intelligence that rivals much larger flagship models while running roughly four times faster — and often at less than half the cost — than other frontier models.
It’s built for agentic and coding work: planning, building, and iterating through long, multi-step tasks. Google pointed to partners using it to automate work that previously took days or weeks, from financial document review to data analysis.
For most businesses, the practical point is access. Gemini 3.5 Flash is now the default model in the Gemini app and in AI Mode in Google Search, so you’re likely already using it. It’s also available through Gemini Enterprise for teams building their own AI agents, and through Google’s developer tools. It’s the same model powering Gemini Spark, the new personal agent in the Gemini app — one of several Workspace-related announcements from I/O.
Gemini 3.5 Pro is coming next
Google confirmed a more powerful Gemini 3.5 Pro is in the works for the heaviest reasoning tasks. It’s already in internal use, and Google expects to roll it out next month. There’s nothing to try yet — but it’s worth knowing the family will have a higher-capability tier soon.
Gemini Omni
Gemini Omni is a separate new model family, and its focus is creating media — starting with video. The first release, Gemini Omni Flash, is rolling out now.
What sets it apart is conversational editing. You can generate a video from a mix of inputs — images, audio, video, and text — and then refine it by describing changes in plain language. Each instruction builds on the last, so characters stay consistent and the scene holds together across edits. Google also built in a better understanding of real-world physics, so generated motion looks more believable.
A couple of practical notes. Omni includes an Avatars feature that can create a digital version of you — using your own voice — to appear in generated videos. And every video Omni produces carries Google’s invisible SynthID watermark, so AI-generated clips can be identified as such.
Gemini Omni Flash is rolling out to Google AI Plus, Pro, and Ultra subscribers through the Gemini app and Google Flow, and at no cost on YouTube Shorts. Access for developers and enterprises through Google’s APIs is expected in the coming weeks.
What this means for your business
The Gemini 3.5 story is mostly about economics. When a capable model runs faster and costs less, more uses become practical — document processing, customer support, internal automation. And because Gemini 3.5 Flash is already the default across Google’s consumer products, the AI inside the tools you use every day just got an upgrade without you doing anything.
Gemini Omni is the one to watch rather than adopt immediately. Conversational video editing makes producing marketing clips, training material, and explainers far more accessible — but for now it’s a consumer-subscription feature, with business and developer access still a few weeks out.
As a Google Partner, we track these releases so you don’t have to. If you’re trying to figure out where Gemini 3.5 or Omni fits into how your business actually works, let’s talk.