Follow ZDNET: Add us as a preferred source on Google.
ZDNET’s key takeaways
- Opera’s Neon AI browser was publicly released on Thursday.
- It comes with a hefty price tag, $19.90/month.
- Gartner recently advised businesses to avoid using AI browsers.
Opera’s AI-powered web browser, Neon, is now available to everyone, but it comes with a hefty price tag of $19.90 per month — probably more than most people are willing to pay for a technology still in its infancy and with unresolved security risks.
Also: Use an AI browser? 5 ways to protect yourself from prompt injections – before it’s too late
Unveiled in October with exclusive waitlist access, Neon is “an experimental browser for AI power users,” Opera wrote in a press release published Thursday. The Norway-based software company is hoping customers will be willing to pay for an AI browser that leverages multiple frontier models and can handle a variety of web tasks. It’s facing ample competition, though: Perplexity and OpenAI have both launched their own free web browsers, called Comet and Atlas, respectively. Meanwhile, Microsoft and Google have been embedding AI features more deeply into Edge and Chrome.
All these new products and upgrades are being sold as an improvement to the traditional web browsing experience, one that largely hands the reins over to AI to handle boring and time-consuming tasks while humans sit back and, in theory, focus their attention on more important and fulfilling tasks. Browsers have become a critical piece of real estate in the ongoing AI gold rush, as they serve as the portal to the internet. If building a chatbot is like owning a ship, then building a web browser is like owning the ocean.
What Neon offers
Opera seems to understand that $19.90/month is a lot to ask of the average consumer, hence the “AI power users” angle.
“Opera Neon is a product for people who like to be the first to the newest AI tech,” Krystian Kolondra, EVP Browsers and European Fintech at Opera, said in a statement. “It’s a rapidly evolving project with significant updates released every week.”
Also: I’ve been testing the top AI browsers – here’s which ones actually impressed me
The biggest perk is that it allows access to a handful of industry-leading AI models, including GPT-5.1, Gemini 3 Pro, and Nano Banana Pro, all of which ordinarily come with their own subscription fees for full and unlimited use.
Neon comes with four specialized AI agents that can handle a specific category of tasks:
- Chat: For having conversations just as you would with any other chatbot.
- Do: To navigate websites and perform actions on your behalf, such as booking a hotel reservation or flight.
- Make: To write code and generate images and other kinds of content.
- Opera deep research agent (or ODRA): To aggregate information, complete with sources, on complex subjects into neatly packaged reports.
Paying the monthly subscription fee also unlocks membership to an exclusive Discord community where users can suggest changes and interact directly with company engineers to “help shape the browser’s roadmap,” according to the company.
Neon can be downloaded here.
Security concerns
Letting AI take the reins on your digital life shouldn’t be done lightly.
When you use an AI browser, you’re granting it permission to autonomously view and handle sensitive personal information and make decisions on your behalf — and this can sometimes go awry. Earlier this month, Gartner published a report advising businesses not to use AI browsers, as it warns that these browsers could share proprietary data with their developers’ cloud servers and are vulnerable to malicious prompt injection attacks, where third parties trick the browsers into carrying out actions that harm users.
Also: Gartner urges businesses to ‘block all AI browsers’ – what’s behind the dire warning
Opera addressed these concerns in an October blog post following Neon’s early access debut: “Just as agentic browsers open up new possibilities in browsing and completing tasks online, so can they be susceptible to new kinds of attacks that exploit vulnerabilities found in AI agents,” the company wrote. The post then outlined steps the company had taken to mitigate those risks, including a “prompt analysis” function that scrutinizes prompts for potential threats.
It also added a caveat: “It is important to acknowledge that due to the non-deterministic nature of AI models, the risk of a successful prompt injection attack cannot be entirely reduced to zero.” In other words, use at your own risk.
