XDA compares ChatGPT, Gemini, and Perplexity for 30 days and chooses the best rival to Claude.

A one-month test conducted by XDA Developers pitted Gemini, ChatGPT, and Perplexity against Claude. in tasks such as research, brainstorming, writing, programming, and document analysis. At the end of the experience, the journalist Parth Shah concluded that the ChatGPT He was the most complete replacement and the only one who truly justified signing the contract. US$ 20The comparison was not based on technical benchmarks or laboratory testing, but on... real everyday use. According to XDAThe idea was to discover whether Claude was still ahead due to merit or if habit was taking over. Over 30 days, each assistant was tested in the same type of routine: researching, organizing ideas, writing, dealing with code, and interpreting documents.

How XDA compared Claude, Gemini, ChatGPT, and Perplexity.

The XDA test sought to answer a simple question: Which AI can best fill Claude's role in everyday life today? Instead of evaluating only isolated answers, Shah took the three contestants to real work tasks and observed how each one behaved throughout the month.

This matters because the competition between AI assistants is no longer just about text quality. Today, factors such as... also weigh in. Integration with other appsConsistency of responses, programming performance, and ease of research using reliable sources. showmetech has already followed the evolution of this market in coverages such as The launch of Gemini 3.5 Flash, focusing on agents and programming. e recent Codex updates.

Gemini was the most impressive feature on the phone.

According to the article, the Gemini: It was the assistant that made the best first impression. Google's redesign was described as more modern, polished, and easier to navigate than previous versions. But the main highlight wasn't visual: it was the integration with Google's own products.

In the test, the author was able to use Gemini with Keep, Tasks, Docs, YouTube Music and WhatsAppIn addition to activating parts of Android through natural language commands, this made the service especially useful on mobile phones, where the assistant seemed more connected to the system and the applications already used daily.

The problem, according to XDA, is that this advantage diminishes on the desktop. On the Mac, the experience was described as merely functional, lacking the same sense of deep integration. In programming, the Gemini 3.5 via Google Antigravity He handled routine tasks and code explanations well, but fell behind Claude when the work required a deeper understanding of complex projects.

Another negative point mentioned was the inconsistency in the tone of the responsesAt times, Gemini was direct and helpful; at other times, it used excessive language or sounded too polite for simple requests. This, in the author's view, prevented it from becoming a complete replacement for Claude.

Perplexity was the best for research and sources.

If Gemini won in mobile integration, the Perplexity was XDA's favorite for researchThe main advantage highlighted was the way the platform presents sources clearly, making it easier to verify information and providing more transparency to the process.

According to Shah, this was especially valuable when researching materials, comparing products, and exploring unfamiliar topics. The feature Perplexity Computer It also drew attention for its ability to navigate websites, gather information, and complete tasks in multiple steps within the browser.

Outside of research, however, the service lost momentum. XDA reports that everyday answers seemed more basic, while writing tasks fell short of what ChatGPT delivered in depth and naturalness. In programming, support was also described as more limited.

Therefore, the conclusion was that Perplexity works very well as a tool for discovery and verification of sourcesbut not as a sole aid for those who want to bring together research, writing, brainstorming, coding, and everyday use in the same place.

ChatGPT was the most complete replacement for Claude.

O ChatGPT It won the test by offering the most balanced experience across desktop, mobile, writing, research, and programming. Unlike Gemini, which impressed more on Android, OpenAI's service was described as consistent across different platforms, with particular emphasis on the native Mac application.

XDA also praised ChatGPT's more open ecosystem, with integration to tools such as Canva, Photoshop, Figma and SpotifyThis made a difference because, according to the author, his workflow doesn't depend on a single ecosystem, like Google's or Microsoft's.

Another strong point was the consistency of the textThe ChatGPT responses were rated as direct, clear, easy to read, and natural, without the excessive dramatization that the author said he felt at times in Gemini.

In the programming section, the highlight was... CodexAccording to Shah, the tool came very close to Claude Code in serious development tasks, including understanding larger projects, fixing multiple files, and building features from broader instructions. XDA still believes that Claude Code may have an advantage in more chaotic scenarios, but states that the gap between the two is now much smaller than it seemed before.

How much do Claude, ChatGPT, Gemini, and Perplexity cost in Brazil?

Since these subscriptions are charged in dollars, the price for subscribers in Brazil varies depending on the exchange rate, IOF (Brazilian tax on financial transactions), and any applicable credit card fees. Using the exchange rate of [insert exchange rate here] as a reference... US$ 1 = R$ 5,11 On June 13, 2026, these are the official values ​​found on the plans' pages. and their approximate equivalents in reais:

Among the plans I was able to verify directly on the official pages, Claude Pro, ChatGPT Plus, and Google AI Pro remain virtually tied in the price range. around R$100 per monthMeanwhile, Google AI Plus remains the cheapest entry-level option. In the case of Perplexity, the official page shown by the user indicates the Pro plan for $17/month when billed annually, placing the service slightly below the $20/month range charged by Claude Pro and ChatGPT Plus.

This shows that the dispute described by XDA happens almost entirely within the same price range: something around R$ 100 per month for Claude Pro, ChatGPT Plus, Google AI Pro and Perplexity Pro, with Google also offering a cheaper entry point into AI PlusIn other words, when the author says that just one of them is "worth US$20," in Brazil this means asking if the tool really justifies a recurring expense close to that amount. 102,15 BRL a month.

Is it worth replacing Claude?

XDA's response was clear: it depends on what you value mostThose who live within the Google ecosystem and use Android extensively can get the most out of Gemini. Those who need transparent research and frequently check sources will find Perplexity a very powerful tool. But for the complete package, the ChatGPT He was the closest thing to a real replacement for Claude.

In closing his analysis, Shah summarizes the dispute directly: after a month using Gemini, ChatGPT, and Perplexity, ChatGPT was the only one that truly seemed worth the $20 subscription fee.For those looking for an AI capable of writing well, researching, programming, and still functioning consistently on desktop and mobile, it was the most convincing winner in the test.

See also:

Text revised by Alexandre Marques in 15 / 06 / 2026.

Source: XDA Developers.