Sundar Pichai Introduces Gemini Spark, Google’s New AI Assistant
Google unveiled Gemini Spark, Gemini 3.5 and Gemini Omni at Google I/O 2026 as Sundar Pichai highlighted the company’s new “agentic Gemini era.”
SociapaHQ
Google announced a major expansion of its artificial intelligence ecosystem at its annual developer conference Google I/O 2026, introducing a range of new AI-powered tools, models and systems including an autonomous AI assistant called Gemini Spark.
The company positioned “agentic AI” as the central theme of this year’s conference, with executives describing the technology as the next stage of AI evolution beyond traditional chatbots.
Speaking during the keynote session near Google’s headquarters in Mountain View, Sundar Pichai said the company had entered its “agentic Gemini era”, NBC News reported.
“We are firmly in our agentic Gemini era,” Pichai said during the event.
Also Read: HDFC Bank Introduces Temporary Hybrid Work Policy Following PM Modi Appeal
AI Agents Arrive:
According to the company, Gemini Spark is designed to proactively assist users by independently performing routine tasks such as sorting through meeting notes, emails and chats before automatically creating summaries, documents and action items.
Unlike standard AI assistants that only respond when prompted, Google said Gemini Spark will continue operating in the cloud even after users close their laptops or lock their phones.
The company added that the AI system will ask users for permission before carrying out “high-stakes” tasks such as sending emails or making purchases.
Select testers will gain access to Gemini Spark beginning Tuesday, while beta access is expected to roll out later for US-based subscribers of Google’s AI Ultra tier. Google also said Gemini Spark will integrate directly into the Google Chrome browser later this summer.
Alongside Gemini Spark, Google also introduced its latest AI model family, Gemini 3.5.
Also Read: Manja Turns The Economic Times Newspaper Into AI-Themed Advertising Canvas
Gemini AI Expands:
The first version rolling out globally is Gemini 3.5 Flash, which the company described as its fastest and strongest coding and agentic AI model so far.
Google said the Flash model is focused on speed and is approximately four times faster than some competing AI systems.
Gemini 3.5 Flash will now become the default model powering the Gemini app and AI Mode within Google Search.
The company also confirmed it is internally testing Gemini 3.5 Pro, which is expected to launch next month.
According to Google, Gemini 3.5 was developed with more advanced safety systems intended to reduce harmful outputs and lower the chances of incorrectly refusing safe user requests.
Google additionally announced a new multimodal AI model called Gemini Omni, designed to generate high-quality videos using prompts that can include text, images, videos and audio inputs.
