- Opera Neon establishes itself as a paid agent browser with a focus on in-depth research and online task automation.
- Debut 1-minute investigation mode with ODRA and work with multiple AI agents in parallel to generate structured reports.
- It integrates the Google Gemini 3 Pro and Nano Banana Pro models, with a model selector that can be switched in the middle of the chat.
- The Do agent now integrates with Google Docs and automates comparisons and redactions, but the service remains in limited access and costs about $20 a month.
After several days of intensive use, Opera Neon leaves a strange feeling: at times it seems like a clear preview of What will web browsing be like in the coming years?, for a while now It feels like a half-baked experiment. which tests the patience of anyone who installs it. Opera's browser is not simply an AI-powered version of its classic product, but a serious attempt to redefine what a browser does when we are no longer the ones clicking on each link.
Neon retains the recognizable foundation of Opera browsers—side messaging integrations, quick access to music services in streamingmultimedia control panel—, but The truly differentiating layer comes with its agentic approach. The idea is that The browser should stop simply answering questions and start acting on behalf of the user.: open pages, compare prices, manage forms or prepare documents while the user focuses on other tasks.
A browser with three main agents and an AI lab underneath
To understand what Opera Neon offers, one must assume that it is not just a browser with an integrated chatbot, but an environment where several different AI agents coexisteach with specific functions. The user moves between them depending on what they need to do, with varying but interesting results.
On one hand there is Chat, the most classic conversational agent, designed to answer questions, Summarize web pages, translate texts, or synthesize informationIts operation is familiar to anyone who has tried other generative AI assistants, and it's useful for quick tasks within the browser itself. However, it suffers from the same problem as many similar models: it occasionally fabricates data or unnecessarily lengthens responses.
Where Opera truly tries to differentiate itself is with DoThe agent responsible for "doing things" on the web. This component can open tabs, browse different sites, fill in fields, and run complete workflows such as searching for a flight, comparing various products, or starting a reservation. Watching Do work is almost hypnotic: It moves around the page, navigates forms, and progresses step by step.The problem is that, to this day, it still does so inconsistently, making mistakes that are difficult to correct on the fly and forcing the user to closely monitor every action.
The third pillar is Make, the creation-oriented agent. Its function is to generate code, small web applications, videos, or other interactive resources directly from the browser. In practical tests, it has been able, for example, to build simple memory games with Spanish vocabulary in a matter of minutes: basic but functional projects that disappear when the tab is closed. It's a kind of integrated "mini-developer," with plenty of room for improvement, but it's geared towards a different type of use than a traditional browser.
This entire system is completed with the so-called Cards, configurable templates of instructions that act as reusable shortcuts promptsThe user can combine these actions—for example, mixing summary and comparison actions or decision-making and follow-up—or create their own to avoid starting from scratch with each interaction. This approach attempts to capture the user's accumulated experience and integrate it into the browser itself, in line with what other agentic tools are exploring.
ODRA and in-depth research in one minute
The big recent development is the incorporation of Opera Deep Research Agent (ODRA), a Specialized agent in advanced investigation that integrates with Chat, Do, and Make to convert the browser into a workplace focused on long reports and analysesInstead of simply returning a short answer, ODRA searches through different sources, cross-references, and generates structured documents with citations.
With the latest update, ODRA launches a “1-minute investigation” mode Designed for those who need something richer than a simple summary, but not a full study that takes several minutes or hours. In this mode, Neon divides the query into multiple subproblems and puts several people to work on them.virtual researchers"in parallel" on the same task. The result is a compact report, with cited sources and a reasonable structure, which aims to be somewhere between a typical chat response and a comprehensive in-depth investigation.
Opera highlights that its deep-search agent scores highly in comparative tests such as DeepResearch Bench, placing it on par with Google and OpenAI solutions for complex analysis tasksBeyond the numbers, the intention is clear: that the browser serves as a useful productivity tool for those who work with a lot of information, not just as a technological showcase.
Model selector and arrival of Gemini 3 Pro and Nano Banana Pro
Another important step in Neon's evolution is the integration of new Google AI models and the ability to choose which one is used at any given timeThe browser now includes a Neon Chat conversation model selectorwhich allows switching between different systems without losing the context of the dialogue.
Among the available options, the following stand out: Google Gemini 3 Pro, geared towards demanding tasks and complex analysesAnd Nano Banana Pro, an image generation and editing model that adds to the browser's visual repertoire. Users can switch between them mid-conversation, preserving their history and session thread, so they can access more powerful options when needed or lighter models for quick queries.
This ability to swap "brain" on the fly seeks to leverage the ecosystem of advanced models without forcing the user to commit to a single option. The approach fits with the idea of Neon as a living laboratory.Opera, which is prepared to integrate AI technologies practically within hours of its announcement, emphasizes that many of these integrations were designed in collaboration with the developer community participating in the early access program.
Agent Do teams up with Google Docs
Among the most frequent requests from early adopters was the integration with cloud-based office toolsThe latest update responds to that demand by allowing Neon Do works directly with Google DocsFrom now on, users can ask the browser to prepare product comparison documents, write drafts, or update existing texts without leaving the tab.
The process is simple: just select the Do agent from the browser menu and add it to the desired instruction. create or edit a Google Docs documentThe agent opens the document, imports the data from the website, adds or removes relevant information, and even changes the file title if requested. In practical terms, this allows for the automation of everything from simple pros and cons lists to more extensive compilations from multiple open pages.
In theory, this type of integration fits very well with Neon's original promise: that the browser assumes and automate repetitive tasks such as collecting data, copying and pasting information, or formatting comparisons, saving time for the researcher. In practice, The experience still requires supervisionThis is especially true when dealing with complex forms, third-party services, or multi-step workflows. Even so, for advanced users who regularly work with shared documents, it's one of the most noticeable improvements in this version.
A paid product in a market where AI is usually free
Beyond its features, Opera Neon stands out for a decision that sets it apart from the rest of the AI browsers on the market: It is a paid subscription serviceAccess to the agentic browser It costs around $19,99 per month and it still is limited to a small number of users within an early access programTo enter, you must register and wait for an invitation.
This strategy clashes head-on with the majority approach in the sector. Currently, giants like Google integrates Gemini into ChromeMicrosoft brings Copilot to multiple products; Perplexity combines its browser with Comet OpenAI offers ChatGPT Atlas as part of its services, often at no additional cost to the end user. The implicit message is that AI in navigation should be ubiquitous and free, at least in its basic functions.
Opera takes a different view: if a browser is going to control tabs, access sites where we are already logged in, manage purchases, or send emailsIt needs an economic model that doesn't depend on monetizing personal data. According to this view, charging a monthly fee would avoid models based on surveillance and invasive advertising, ensuring that the customer is the user and not the advertising intermediaries, and helping to Protect your privacy.
Neon's technical architecture points in that direction, with a hybrid system where the most sensitive tasks are executed locally without sending passwords to the cloud, while other processes rely on remote servers. It's a strategy that It comes at a complex time.This comes amid a saturation of AI services and with users increasingly tired of new subscriptions, but it raises a relevant debate about who is in control of the future agentic web.
Opera Neon within the Opera browser ecosystem
Neon does not replace the company's main browser nor to the rest of the brand's products. Opera maintains its traditional offering, with Opera One as the flagship For those seeking a pleasant and versatile browsing experience, Opera GX is geared towards the public gamer y Opera Air with a more minimalist approach, and alternatives such as Sidekick browserAll of them include free AI solutions that are independent of specific language models.
In that context, Neon positions itself as The experimental option for advanced users who want to influence the future of browsingOpera openly defines it as a "testing ground" in which to rapidly introduce the latest AI technologies, adjusting the experience based on feedback from a relatively small but very active community. Hence, features as mature as one would expect in a commercial product coexist with others that still exhibit erratic behavior.
The Norwegian company boasts some 300 million users across all its browsers, but it's aware that not everyone is looking for the same thing. Instead of a single solution for all users, it offers a family of products where Neon occupies a key position. the riskiest and most speculative space, designed for those who accept living with flaws in exchange for being one step ahead in navigation trends.
Between technological fascination and the seams of a beta face
My experience with Opera Neon reflects this duality. On the one hand, it's stimulating to see a browser attempt more than just embedding a chat box in a sidebar. The way Do moves through pages, how ODRA distributes a complex query among several agents The possibility of switching between Google models to better leverage their strengths paints a picture of a future where many online bureaucratic tasks will be delegable.
On the other hand, the system still retains an openly experimental character. Errors in interpretation of Do, overly long responses from Chat, unpolished examples of Cards, and the need to manually correct actions that the agent doesn't fully understand all contribute to this. The promise of "the browser that works for you" has not yet been consistently fulfilled.Neon can save time in certain very specific cases, but it also wastes time when it forces processes to be repeated due to agent failures.
In this context, the fee of around $20 per month puts the product in a precarious position compared to free alternatives or those included within other services. The audience it might best suit today is the so-called power users: people who spend a good part of the day comparing information, preparing reports, or building small tools and that they are willing to pay for an advance of what is to come, even assuming imperfections.
Today, Opera Neon presents itself as a intriguing agentive browser And still immature, a paid “testing ground” that offers real advancements in task automation, rapid research, and integration with advanced Google models, but requires tolerating a fair amount of friction. For the average European user, who already has established browsers and free AI features, its offering is more of an invitation to participate in the experimental phase of the next generation of browsers than an immediate replacement for the tools they use daily.
I am a technology enthusiast who has turned his "geek" interests into a profession. I have spent more than 10 years of my life using cutting-edge technology and tinkering with all kinds of programs out of pure curiosity. Now I have specialized in computer technology and video games. This is because for more than 5 years I have been writing for various websites on technology and video games, creating articles that seek to give you the information you need in a language that is understandable to everyone.
If you have any questions, my knowledge ranges from everything related to the Windows operating system as well as Android for mobile phones. And my commitment is to you, I am always willing to spend a few minutes and help you resolve any questions you may have in this internet world.