Opera Neon: Inside the $20 "Agentic" Browser Everyone Is Talking About

The Bottom Line: Built for Power Users, Not for the Casual User

Opera Neon's start page with agentic AI commands for research and tasks. Image: Opera

Quick Verdict

The 2025 re-launch of Opera Neon marks the first mainstream attempt to monetize an "Agentic Browser"—a tool that actively browses the web for you. With a $19.90/month price tag, it is clearly positioned for power users requiring access to premium models like Gemini 3 Pro and GPT-5.1 within a single, unified interface. While we haven't run our own long-term stress tests yet, the specs suggest this is a massive productivity unlock for developers and researchers, though likely overkill for the casual user.

Today, Opera finally launched its sophisticated AI Agent browser to the public, designed to automate user-assigned tasks. The company has branded this new AI-powered tool as Neon Browser, placing it behind a premium paywall of $19.90. After months of beta testing which began in May 2025, the company has opened access to everyone starting today.

Although this new browser is not a brand-new concept, we have already seen similar iterations from competitors like OpenAI’s Atlas, Perplexity Comet, and The Browser Company’s Dia, Neon is fundamentally different from standard browsers like Google Chrome, Microsoft Edge, or even the company’s own Opera GX and Opera One.

Is the $20 Price Tag Justifiable?

The core of Neon isn't just for viewing the web; it is for doing things on the web. It integrates an AI chatbot directly into the interface with deep context awareness. The question on everyone's mind is: Is the $20 price tag worth it?

The short answer is yes, if you are the right user. The browser utilizes top-tier Large Language Models (LLMs) for its agentic features, specifically Google Gemini 3 Pro and GPT-5.1. Subscribing to these services separately is significantly costlier, but the new Neon browser integrates both under one subscription. Furthermore, the browser uses local processing to perform actions on websites (like booking, filing, and sorting) securely, addressing major privacy concerns.

The Three Pillars: Chat, Do, and Make

Opera has structured Neon around three specific modes that aim to replace the traditional tab-browsing experience.

1. Neon Chat (The Context Engine)

Most browsers have a chat sidebar, but Neon differs by integrating the AI directly into the rendering engine.

• The Feature: It can "see" the DOM (Document Object Model) of your open tabs.
• The Benefit: This theoretically reduces AI hallucinations because the agent is reading the actual source code of the page you are viewing, rather than just guessing based on surface text.

2. Neon Do (The Agent)

This is the feature driving the "Agentic" hype. Neon Do is designed to execute multi-step workflows autonomously.

• How it works: Instead of searching for "flights to London," you command the browser: "Find the cheapest direct flight to London on Expedia for next Tuesday and add it to my calendar."
• Privacy Focus: Opera emphasizes that Neon Do operates locally using your existing login cookies. Your passwords are not sent to an external OpenAI or Google server; the "robot" clicking the buttons lives inside your laptop.

3. Neon Make (The Local Builder)

For developers and data analysts, this is the killer app. Neon Make allows the browser to generate temporary, local web apps.

Use Case: You can ask Neon to scrape data from 10 open tabs and "Make a comparison dashboard with a bar chart." The browser writes the HTML/JS instantly and renders it as a new interactive tab.

Deep Dive: Analyzing the "Agentic" Capabilities

Since the public waitlist dropped today, we are analyzing the capability set based on Opera's technical documentation and early "Founders" community demos.

The "Neon Do" Workflow and Session Persistence

The biggest shift here is Session Persistence.

• The Old Way: Traditional AI agents usually fail at "logging in" or handling CAPTCHAs.
• The Neon Way: Because the agent lives in the browser, if you are already logged into Amazon or Jira, the agent is too. This bypasses the biggest hurdle for automated web agents: authentication.

The Model Economy

Switching between Gemini 3 Pro and GPT-4.1 models in Opera Neon AI. Image: Opera

For $20/month, the math is compelling for heavy users.

Standalone Cost: A Gemini Advanced subscription is ~$20. ChatGPT Plus is $20.

Neon Bundle: Opera claims to offer "Pro" level access to both models (likely via API credits capped at a high usage limit). If you regularly switch between modelsusing Gemini for its large context window and GPT for its reasoning—this subscription actually saves you money.

The Privacy Argument: Local Agent Processing

The strongest selling point for the Opera Neon browser might be its Local Agent Processing. Many users are hesitant to use cloud agents (like Rabbit or others) because they don't want to share login credentials. By keeping the "Do" agent local, Opera Neon solves the trust gap for users who want automation without security risks.

The $20 Question: Is It Worth It?

This is a premium tool for a specific audience. Here is who should consider the upgrade:

User Type	Verdict	Why?
Developers	Strong Buy	"Neon Make" can prototype UI changes and clean data faster than manual coding.
Researchers	Buy	The context window allowing cross-analysis of 20+ tabs is a major time saver.
General Users	Skip	If you primarily browse social media and news, the free AI in Chrome or Edge is sufficient.

Final Thoughts

Opera Neon is currently a "Concept Car" for the future of the web. It demands you change how you browse, moving from "searching" to "commanding."

Next Step: Opera has removed the waitlist as of this morning. You can download the trial version to see if the "Agentic" workflow fits your style before committing to the subscription. If you are hesitant about the price, stick to the free Opera One R2, which includes basic Aria AI features without the agentic capabilities.

Note: This analysis is based on the technical release specifications and feature demos available as of December 2025.

Source: Image1, Image2