Google is pushing the boundaries of its Gemini project, testing fresh modes that go well beyond the recently spotted gem pix image-editing tool tied to the nano-banana model. The updates in the mode selector point to three initiatives. The standout is Agent Mode, introduced at Google I/O, which is built to handle autonomous exploration, planning, and task execution. This puts Gemini in line with workflows you might recognise from tools like ChatGPT, shifting it from a simple prompt-response service to a more capable multi‑step assistant.
Alongside Agent Mode, Gemini Go and Immersive View are in the mix. Gemini Go is all about exploring ideas together – it may well tie into Canvas, the platform already known for prototyping and sketching, although whether it stands alone or integrates remains to be seen. Immersive View, on the other hand, promises visual answers to your questions, perhaps evolving from Google’s Video Overviews or venturing into new image-generation territory. Even if not every mode launches as its own dropdown, Agent Mode’s unique icon suggests it’s set to become a permanent feature, while the others remain experimental as they find their place.
With these enhancements, Google is clearly reimagining Gemini as more than just a chat assistant, opting instead for an integrated workspace that supports creative thinking, research, and visual explanation.
