All Eyes on Google I/O: Rumors Swirl Around Massive Gemini 3.5 Reveal

With Google I/O 2026 just days away, speculation is intensifying that Google could unveil its next major artificial intelligence model: Gemini 3.5.

While Google has not officially confirmed the existence of Gemini 3.5, a growing number of leaks, insider hints, backend discoveries, and industry reports have fueled expectations that the company is preparing a major AI announcement for its annual developer conference in May.

Why Analysts Believe Gemini 3.5 Could Debut at I/O

First, Google executives have recently teased that a “new Gemini model” is arriving “very, very soon,” according to reports following comments from Google Cloud leadership.

At the same time, developers and AI researchers claim they have discovered references to “Gemini 3.5” hidden in Google API variables and internal testing environments. Some reports describe an experimental version carrying the codename “Snow Bunny,” allegedly capable of significantly stronger reasoning and large-scale code generation.

The timing also aligns with Google’s historical launch strategy. Google frequently uses I/O to introduce major AI upgrades, Android announcements, and next-generation developer tools. Industry watchers say the company is under increasing pressure to showcase rapid AI progress amid fierce competition from OpenAI, Anthropic, and Meta.

Rumored Features of Gemini 3.5

Though none of the reported capabilities are confirmed, leaks and speculation suggest Gemini 3.5 could focus on several key upgrades:

Improved reasoning and “Deep Think” capabilities
More advanced autonomous AI agents
Longer memory and context windows
Enhanced multimodal understanding
Stronger software engineering and coding performance
Unified video, audio, and image generation tools

Some rumors also point toward a broader AI initiative internally referred to as “Remy,” described as a proactive AI assistant capable of taking actions on behalf of users instead of simply responding to prompts. Multiple reports suggest Google may preview these agentic AI features during I/O.

AI Community Watching Closely

The AI community has become increasingly convinced that Google is preparing a major reveal.

Prediction markets and online forums have seen rising discussion around the odds of a Gemini 3.5 launch before the end of May. Meanwhile, rumors circulating on Reddit and developer communities claim Google may also unveil related projects such as “Gemini Omni,” a multimodal video-generation platform designed to compete with OpenAI’s Sora.

Developers have also observed what they describe as unusual activity around Gemini Flash model updates and deprecations — often seen as signs that Google is preparing for a broader model transition.

What Google I/O 2026 Could Mean for the AI Race

If Gemini 3.5 is announced, analysts believe it would represent Google’s most aggressive attempt yet to reclaim leadership in consumer AI.

Over the past year, OpenAI’s GPT ecosystem and Anthropic’s Claude models have dominated many developer and enterprise conversations. Google, despite its deep AI research roots through DeepMind, has faced criticism for slower product rollouts and fragmented AI experiences.

A successful Gemini 3.5 launch could change that narrative.

“This year’s I/O feels unusually important,” one AI researcher posted online ahead of the conference. “Google needs to prove it can move as fast as the rest of the industry.”

Still No Official Confirmation

Despite mounting evidence and speculation, Google has remained silent on whether Gemini 3.5 actually exists — or whether the company may instead pivot directly to Gemini 4 branding.

For now, the rumors remain exactly that: rumors.

But with Google I/O 2026 approaching rapidly, the tech world is watching closely to see whether Google is about to unveil its next big leap in artificial intelligence.