At the recent Google I/O developers conference, Google introduced its latest Gemini AI models, which are specifically engineered to function as autonomous agents. The rollout of Gemini 3.5 Flash marks a shift in how AI can assist users by not just answering queries but actively performing tasks. This evolution comes at a time when the demand for more capable AI solutions is rising amid intense competition from other leading AI firms.
New Model Launches
Gemini 3.5 Flash is now available to all users. According to Google CEO Sundar Pichai, it is designed to operate efficiently while delivering tasks at a lower cost compared to competing models. The model can run multiple agents simultaneously, allowing it to manage extensive, long-term projects independently. Koray Kavukcuoglu, Google's chief AI architect, emphasized that this capability enables Gemini 3.5 Flash to conduct entire coding research projects over extended sessions.
Anticipation is building for the forthcoming Gemini 3.5 Pro, currently undergoing internal testing, which is expected to offer even more advanced functionalities. This model is positioned as a direct competitor to other leading AI models, including OpenAI's GPT-5.5 and Anthropic's Claude Opus 4.7, with Google claiming its offerings are superior in speed and task efficiency.
Introducing Gemini Spark
Accompanying the new models is Gemini Spark, an always-on AI assistant aimed at competing with existing platforms like OpenClaw. This autonomous agent can operate within Google Workspace applications such as Gmail and Google Docs, executing tasks on behalf of users. Josh Woodward, vice president for Google Labs, described Spark’s functionality as intuitive, stating, "When you use it, it almost feels like you're tossing things over your shoulder, Spark's catching them and getting the job done." Initially, access to Spark will be limited to a select group of trusted users, with a broader beta release for AI Ultra subscribers expected soon.
As more integrations are planned over the summer, Spark’s capabilities will expand, enabling it to interact with a wider array of third-party applications through Chrome. This expansion reflects Google’s commitment to enhancing user productivity through AI.
Revamping Antigravity
To strengthen its coding capabilities, Google has reimagined its Antigravity platform, positioning it as an agent-first development tool. Kavukcuoglu noted that this platform will serve not just for coding but also for managing teams of autonomous AI agents. The revamped Antigravity will feature a standalone desktop application that supports various tasks, from website creation to product planning, showcasing its versatility in AI development.
Enhancements in the Gemini App
The Gemini app is also undergoing a significant redesign, integrating the new models along with features aimed at improving user interaction. The updated interface, termed Neural Expressive, introduces new animations and haptic feedback, enhancing the tactile experience of using the app. A new feature called the Daily Brief will provide users with personalized summaries based on their inbox and calendar data, allowing for a tailored experience that highlights the power of these autonomous agents.
With these advancements, Google is not only enhancing its AI offerings but also establishing a more comprehensive ecosystem for users to engage with autonomous technology. As AI continues to evolve, Google’s focus on agent-driven solutions positions it to compete aggressively in the market, paving the way for future developments in AI capabilities and user engagement.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
