I evaluated Gemini Advanced, ChatGPT, and Copilot Pro. Here’s how they compared in search performance.

      With AI chatbots now integrated into search engines, browsers, and desktops, it’s easy to think they all function similarly. However, when it comes to providing useful search results, some excel beyond the others.

      I decided to compare Gemini Advanced, ChatGPT, and Copilot Pro directly to determine which one delivers answers more quickly and accurately. These are the paid versions, each promising real-time web access, enhanced contextual understanding, and fewer inaccuracies.

      I presented each AI with the same prompts—from current events to in-depth research questions—and evaluated them on five criteria: accuracy, depth, follow-up capability, errors, and usability. Here’s how they compared.

      **Test 1: Accuracy and real-time information**

      To begin, I posed a current events question requiring real-time knowledge rather than general facts: “Who won the latest NBA playoff game?” Gemini Advanced provided just a scoreboard displaying the teams and final scores, without any additional context, highlights, or player statistics. It also referenced scores from May 10, two days prior, which is not ideal for a real-time inquiry.

      ChatGPT Plus, in contrast, offered a more detailed response, including specifics like the Timberwolves leading the series 3-1 against the Warriors. It mentioned that Julius Randle and Anthony Edwards scored a combined total of 61 points—31 from Randle and 30 from Edwards. Moreover, it included functional source links beneath each paragraph, allowing for easy verification of information. I appreciated how hovering over the source link highlighted the text drawn from that reference. My only critique is that it buried the answer in excessive details; a brief summary at the beginning would have been beneficial.

      Copilot Pro, on the other hand, provided a succinct answer right away and inquired if I wanted more information. For this round, I have to give credit to Copilot Pro—it delivered the direct answer effectively and even suggested a follow-up.

      **Test 2: Depth of response**

      For the second test, I asked a broader question that needed more than just a quick fact: How can I create a strong password? Gemini Advanced offered more tips than ChatGPT and included source links for each suggestion for easy validation. Its longer sentences made the response easier to read without excessive scrolling, distinguishing it from ChatGPT, which provided fewer tips and lacked source links. However, ChatGPT did ask if the conversation was helpful, a point where Gemini fell short.

      Copilot Pro contributed less information and did not provide any source links. Nevertheless, it suggested a few relevant follow-up questions, such as: Why is a strong password crucial for security? Can you provide an example of a strong password? How does a password manager maintain my security? I also enjoyed the added emojis next to each tip.

      **Test 3: Follow-up flexibility**

      In this test, I asked each AI a follow-up query to evaluate how well it maintained context and grasped my intent. I followed up with, “Can you explain why using personal information in passwords is bad?”

      ChatGPT provided three main points, several additional security tips, and a concluding summary. Copilot Pro offered three tips along with a few sentences on maintaining safety. However, Gemini was the only one that neglected to provide specific safety tips at the end. It presented a few more reasons why using personal information is detrimental and added some extra information.

      I have to acknowledge that Copilot Pro and ChatGPT excelled in this round, while Gemini has room for improvement. None of the three included source links here, which felt like a missed opportunity.

      **Test 4: Mistakes and hallucinations**

      One significant concern with any AI assistant is its propensity to confidently present incorrect information. They can provide statements that range from humorous to concerning. I gave each chatbot a series of factual prompts to assess accuracy and whether they acknowledged uncertainties, and all performed exceedingly well in this regard.

      I started with a straightforward question: when was Microsoft founded? Gemini Advanced responded simply, “Microsoft was founded in 1975.” ChatGPT provided more detail, stating, “Microsoft was founded on April 4, 1975, by Bill Gates and Paul Allen.” Copilot Pro gave an even longer response: “Microsoft was founded on April 4, 1975, by Bill Gates and Paul Allen in Albuquerque, New Mexico, USA. It started as a small software company and quickly grew into one of the world’s largest and most influential tech companies. Quite the success story, right?” I appreciated how Copilot found a balance, offering enough context without overwhelming me and suggesting three clickable follow-up questions. I have to admit, I favored Copilot Pro's response.

      Next, I asked all three AI assistants, “Which is the best AI assistant available?” Gemini provided a comprehensive overview of the leading AI assistants, including a quick summary of their functionalities. I particularly liked the part where it indicated which assistant might be the ideal option depending on user priorities, such as choosing Gemini for certain features while opting for Chat