Google is escalating the AI arms race with two new Gemini models and a personal agent, directly targeting the territory of rivals like OpenAI and Anthropic.
Back
Google unveils 2 new Gemini models to challenge OpenAI
Google (GOOGL) unveiled Gemini 3.5 Flash, a new AI model designed for speed and efficiency, alongside Gemini Spark, a personal AI agent capable of executing background tasks, in a major push to reclaim leadership in the generative AI space.
"It's your personal AI agent that helps you navigate your digital life, taking action on your behalf and under your direction," Alphabet CEO Sundar Pichai said during a briefing, highlighting Spark's ability to run on dedicated cloud infrastructure without needing a user's device to be active.
According to the company, Gemini 3.5 Flash, which now powers Google's core search engine and Gemini app, is designed to outperform other frontier models in speed, running four times faster than the industry's quickest alternatives. Spark integrates natively with Google Workspace and uses Chrome for web tasks, with a wider beta release planned for AI Ultra subscribers at $100 a month. Google also introduced Gemini Omni, a multimodal model for video generation.
The announcements position Google to compete more aggressively with Anthropic and OpenAI, which have gained significant traction in the enterprise AI market. While Google's cloud division saw 63% year-over-year growth to $20.03 billion in the latest quarter, its overall share of paid AI business subscriptions stood at just 4.5% in April, compared to over 32% for each of its main rivals, according to data from Ramp.
Spark Agent Aims for Autonomy
Gemini Spark represents Google's most direct entry into the field of autonomous AI agents. The agent is designed to operate continuously in the background on Google's cloud, allowing it to manage long-running tasks like monitoring email inboxes for customer questions or compiling progress updates from various documents without requiring the user's machine to be on. Users can monitor the agent's activities from their phones through a new feature called Android Halo. This focus on autonomous, long-horizon tasks appears to be a direct response to similar agentic capabilities recently launched by competitors.
Flash 3.5 and Search Overhaul
The new Gemini 3.5 Flash model serves as the engine for a significant overhaul of Google's flagship search product. The company is introducing an "Intelligent Search Box" that expands to handle more complex, conversational queries and can accept images, files, and video as inputs. Google claims Flash 3.5's speed and performance, which it says outperforms Gemini 3.1 Pro on benchmarks like Terminal-Bench 2.1, are what make these enhanced agentic search features possible. The focus is on turning Search into a central hub for completing tasks, not just finding information.
The updates reflect a significant strategic bet for Google, which plans to spend between $180 billion and $190 billion this year on AI infrastructure and chips. For investors, the key question is whether these new capabilities can translate into a larger share of the enterprise AI market, where Anthropic and OpenAI currently dominate. While Google's Gemini Enterprise product saw 40% quarterly growth in paid users, the company is still playing catch-up in a market that is moving at a rapid pace.
This article is for informational purposes only and does not constitute investment advice.
[1] Summary of Key Points from Google I/O Conference[2] Google targets AI agents and video generation with Gemini 3.5 Flash and Omni[3] Google is making its biggest change to the search bar in years[4] Google launches Gemini Spark personal AI agent at I/O 2026[5] Google Shoves More AI Into Search, Including A Dynamic Search Box And Agentic...
