Google unveils AI model Gemini 3.5 and AI agent Gemini Spark

Google It launches the latest version of Gemini and a new AI model designed to simulate the physical world; The search giant is racing to keep up with model development while also offering more agency services to its massive user base.

The company made the announcements at the annual Google I/O developer conference on Tuesday, gaining an audience for new product launches at a time when the market is focused on soaring valuations of OpenAI and Anthropic as they prepare for IPOs this year.

The centerpiece of Google’s AI strategy is Gemini, its family of models and tools. The company is showcasing the Gemini 3.5 Flash, a lighter addition to its package that offers superior capabilities at half, or in some cases close to a third, the price of similar flagship models, according to CEO Sundar Pichai.

In a press conference with reporters ahead of Tuesday’s event, Pichai said the Gemini 3.5 Flash is “extraordinarily fast.” The company said 3.5 Flash will now be the default model for the Gemini app and AI mode in global searches.

“You no longer have to trade quality for latency,” Google said in a blog post. The company said Gemini 3.5 Flash strengthens cybersecurity defenses, making it “less likely to generate malicious content and accidentally refuse to answer secure queries.”

Google said its heavier version, Gemini 3.5 Pro, is in use internally but won’t be ready for wider distribution until next month.

On the agency AI front, Google announced Gemini Spark, a new general-purpose AI agent in its Gemini app that can reason between information in connected applications. Google said it wants to help users navigate their digital lives by “taking action on your behalf while under your direction.” Gemini Spark is in beta and will be available first to trusted testers and Google AI Ultra subscribers starting next week.

As more internet users turn to chatbots, Google is trying to convince traditional search users that it can be trusted to help them with tasks that require minimal input. Following the company’s skyrocketing capital spending, Wall Street is looking to Google to show it can create deeper integrations across its products, and agencies could be one way to do that.

The prospects for AI companies continue to grow, especially in light of Anthropic’s recently released Mythos model, which is said to be powerful enough to find thousands of previously unknown vulnerabilities in the world’s software infrastructure.

Google’s AI portfolio now includes Omni, a world model designed to simulate physical environments and predict what will happen next based on the user’s actions. Earth models are often used in robotics and games and have been extensively researched by DeepMind over the years.

Omni will work in Flash, Gemini App, Google Flow and YouTube short videos, and will support image and audio, the company said in a separate blog post, adding that users will be able to use Omni to edit videos and create more realistic images.

“Take a video you’ve taken and ask Omni to change what’s going on,” the post says. AI can “edit the action, add new characters or objects.”

WRISTWATCH: What to expect from Google I/O?