Google is preparing major updates to Gemini web ahead of the upcoming Google I/O event, with several new features surfacing through code discoveries and recent announcements. The Gemini toolbox, known as Agents, is expanding to include options like Memory (Teamfood), Veo 2 for video generation, and Image, alongside existing tools such as Canvas, Deep Research, and Search. While some of these features were previously available, their addition to the Agents toolbox suggests a move toward a unified, multi-modal workspace for users.
A notable new element is the MMGEN Discovery Card, which appears to be a discovery pop-up for multi-modal generation, potentially signaling a broader rollout of multi-modal capabilities. This aligns with recent rumors about the introduction of Gemini Ultra, a new subscription tier that would consolidate advanced features—such as video and image generation—under one plan. Evidence from code strings points to rate limits for video generation, with prompts to upgrade to Gemini Ultra for higher usage, indicating a tiered subscription model similar to what is seen with Gemini Advanced and the rumored Pro tier.
The Deep Research feature is also set to receive a significant upgrade, allowing users to upload files—including images, code, and documents—for more tailored research insights. This has been a highly requested feature and is expected to roll out in the coming months. The ability to ground research against user-uploaded files would make Deep Research much more versatile for professionals and researchers.
Google’s strategy appears focused on consolidating its Gemini offerings, providing more granular subscription options, and keeping pace with competitors like OpenAI. The company is likely to announce these updates at Google I/O, positioning Gemini Ultra as a direct response to anticipated multi-modal systems from rivals. These changes reflect Google’s ongoing shift toward subscription-based AI services, aiming to serve both casual users and professionals needing advanced capabilities.
