AI browser Opera Neon gains new tools and Gemini 3 and Nano Banana support
Opera has released a major update for Opera Neon, the experimental agentic browser it launched two months ago. The update adds a new one minute mode to the Opera Deep Research Agent (ODRA), introduces Google’s Gemini 3 Pro and Nano Banana Pro models, and lets users choose which AI model to use in conversations. It also expands Neon’s agentic features to support tasks inside Google Docs.
Neon is designed for testing AI features and exploring agentic browsing, where AI tools don’t just answer queries but can also complete multi step tasks.
“Opera Neon is our experimental playground that allows us to tap into the most cutting-edge AI technologies that exist on the web. We’re developing this product together with our community at an extreme speed. The way Opera Neon’s AI engine is set up allows us to integrate the latest technologies barely hours after they become available,” said Krystian Kolondra, EVP Browsers at Opera.
The main addition in this latest release is the improved Opera Deep Research Agent. The tool currently produces structured, citation backed reports by cross checking information from multiple sources, and the new one minute mode shortens this process.
ODRA splits tasks into parallel threads, similar to assigning different researchers to different parts of a job. The new mode is somewhere between a basic chat response and a full deep research request, aimed at users who want more than a basic answer but don’t need a lengthy report.
Opera Neon Chat
The update also introduces a model selector for Neon Chat, allowing users to choose which model they want for each conversation, including Google’s newly released Gemini 3 Pro and the Nano Banana Pro image generation and editing tool. Conversations can continue after switching models with the chat history fully remembered.
Another new feature in this release, which Opera says was requested by the Neon community, is support for Google Docs which has been added to Neon Do, the agent that performs user requested tasks. Neon users will now be able to ask the browser to create a Google Doc, add information to an existing one, or make changes to a document's contents. Neon Do will also be able to rename documents and look up info from the web as part of a task.
Opera Neon is currently in early access and only available to invited users. You can find out more, and register your interest here.
What do you think about experimental browsers built around AI agents? Tell us in the comments.