{"id":32689,"date":"2025-05-26T13:09:29","date_gmt":"2025-05-26T13:09:29","guid":{"rendered":"https:\/\/www.mon-agent-ia.fr\/blog\/?p=32689"},"modified":"2025-05-26T13:09:30","modified_gmt":"2025-05-26T13:09:30","slug":"chatgpt-gemini-and-perplexity-the-competition-for-deep-search-innovation","status":"publish","type":"post","link":"https:\/\/www.mon-agent-ia.fr\/blog\/en\/chatgpt-gemini-and-perplexity-the-competition-for-deep-search-innovation\/","title":{"rendered":"ChatGPT, Gemini, and Perplexity: The Competition for Deep Search Innovation"},"content":{"rendered":"

In a rapidly changing technological landscape, where every player is seeking to take center stage, innovations in deep research have become a veritable battleground. In 2025, companies like OpenAI, Google, and Perplexity AI are competing to offer advanced features, opening up new perspectives for users. Users are quickly realizing that response quality and speed are essential in our hyperconnected world, highlighting the solutions these companies offer. Google’s Advances with Gemini Deep Research<\/strong>In December 2024, Google launched its Gemini app, integrating a deep search feature that has since attracted many subscribers. Available for \u20ac22 per month, this feature offers the ability to obtain answers via the Gemini 1.5 Pro model, but with a wait time that can range from 5 to 10 minutes per query. This approach is clearly distinct from that of Gemini Flash Thinking<\/strong>, a faster solution that combines reasoning and multi-source research, particularly through tools like YouTube and Google Maps. Users interested in how Gemini Deep Research works will notice some innovative elements:<\/strong> An interactive search plan: the user can edit it as needed.<\/p>\n\n

A progress sidebar: providing URLs viewed in real time.<\/h2>\n\n

The ability to integrate with multiple information sources for more robust results. Competition with Perplexity AI<\/strong> On the opposite front, Perplexity AI<\/strong>also introduced its Deep Research feature. Currently in the pilot phase, Perplexity offers a free version that allows up to 5 queries per day. For Pro subscribers, the limit increases to 500 queries, extending access to accurate and relevant information. This feature stands out for its ambitious promise of reducing search time. In 2 to 4 minutes, Perplexity Deep Research could accomplish work that would take a human expert hours. Multiple models are available, including GPT-4 Omni and Claude 3, allowing for great flexibility depending on user preferences. Features<\/strong>Google Gemini Deep Research<\/p>\n\n

Perplexity Deep Research<\/p>\n\n

  • Monthly Cost<\/li>
  • \u20ac22<\/li>
  • Free \/ Pro Version \u20ac20<\/li><\/ul>\n\n

    Response Time<\/h3>\n\n

    5 to 10 minutes 2 to 4 minutes<\/strong> Query Quotas<\/p>\n\n

    Limited (unspecified)<\/p>\n\n

    \n\n\n\n\n\n\n
    5 queries (free) \/ 500 queries (Pro)<\/th>\nOpenAI’s Leading Role in the Deep Research Arena<\/th>\nOpenAI, whose commitment to deep research is crucial, has also been offering its own version of this service since February 2025. With a subscription cost of \u20ac229 per month for ChatGPT Pro, OpenAI provides its users with up to 100 daily queries.<\/th>\n<\/tr>\n<\/thead>\n
    This solution is based on a fine-tuned version of GPT-3, incorporating reinforcement learning techniques based on web browsing. The promise is clear: to perform in 5 to 30 minutes what would take a traditional human researcher hours. Thus, OpenAI positions itself as one of the pioneers in this field, and the development of its technology continues to evolve at a rapid pace.<\/td>\nTo remain competitive, OpenAI plans to expand its Deep Search functionality to other paid plans, thus increasing accessibility while promising more generous quotas in the future.<\/td>\nKey Elements of the Deep Search Feature<\/td>\n<\/tr>\n
    Users can expect several features integrated into the OpenAI service:<\/td>\nUse of advanced methods to improve the relevance of results.<\/td>\nAbility to perform more efficient searches thanks to an advanced machine learning model.<\/td>\n<\/tr>\n
    Expected expansions to other data sources in the future.<\/td>\nPerformance Review on Humanity’s Last Exam<\/td>\nA key indicator of the effectiveness of deep search services is the Humanity’s Last Exam benchmark, launched by Scale AI and the Center for AI Safety. With approximately 3,000 questions on a multitude of topics, it serves as a benchmark for judging model accuracy. The results of this benchmark show that OpenAI’s Deep Search mode achieved an accuracy rate of 26.6%, while Perplexity’s achieved 21.1%. These figures not only reveal the current limitations of each solution, but also highlight the importance of competition to constantly improve the services offered to users.<\/td>\n<\/tr>\n<\/tbody>\n<\/table><\/figure>\n\n

    Society<\/h2>\n\n

    Accuracy Rate % on Humanity’s Last Exam<\/strong>OpenAI<\/p>\n\n

    26.6<\/p>\n\n

    Perplexity AI <\/p>\n\n

    21.1<\/h3>\n\n

    Future Expectations and Impact on Innovation<\/p>\n\n

    • In the future, companies like<\/li>
    • Google<\/li>
    • ,<\/li><\/ul>\n\n

      OpenAI<\/h2>\n\n

      , and Perplexity AI<\/strong>will play a leading role in defining deep search standards. These improvements will influence not only how users conduct their searches, but also industry standards.<\/p>\n\n

      Another company,<\/p>\n\n

      \n\n\n\n\n\n
      Anthropic<\/th>\n, is experimenting with similar approaches, adding an extra dimension to the competition. Advances in language models and artificial intelligence are driving a quest for precision and efficiency, driving innovation at a breakneck pace.<\/th>\n<\/tr>\n<\/thead>\n
      Users will benefit from this competition, as they will see an improvement in the quality of responses and the speed with which these results are delivered.<\/td>\nWhat impact does this have on the end user?<\/td>\n<\/tr>\n
      With these developments, several points will be crucial for end users:<\/td>\nEasier access to relevant and accurate information.<\/td>\n<\/tr>\n<\/tbody>\n<\/table><\/figure>\n\n

      Reduced response time, making searches more efficient.<\/h2>\n\n

      Ability to explore multiple sources in record time. Conclusion on innovation in deep search<\/strong>As competition intensifies among large technology companies, the continued development of deep search is transforming the way users interact with artificial intelligence systems. Each advancement and improvement promised by players like Google, OpenAI, and Perplexity AI defines a new playing field for information retrieval, with increasing expectations for efficiency and relevance. Ultimately, it’s essential to monitor these innovations, as they will not only determine the commercial success of the companies in question, but will also shape user interactions with technology in the years to come. <\/strong> <\/strong> <\/p>\n\n

      <\/strong><\/p>\n\n

      <\/p>\n\n

      <\/h3>\n\n

      <\/p>\n\n

      • <\/li>
      • <\/li>
      • <\/li><\/ul>\n\n

        <\/h2>\n\n

        <\/strong> <\/strong> <\/strong> <\/p>\n\n

        <\/p>\n\n\n","protected":false},"excerpt":{"rendered":"

        In a rapidly changing technological landscape, where every player is seeking to take center stage, innovations in deep research have become a veritable battleground. In 2025, companies like OpenAI, Google, and Perplexity AI are competing to offer advanced features, opening up new perspectives for users. Users are quickly realizing that response quality and speed are […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1398],"tags":[1968,2872,19743,813,31450],"class_list":["post-32689","post","type-post","status-publish","format-standard","hentry","category-news-ai-en","tag-chatgpt-en","tag-gemini-en","tag-in-depth-research-en","tag-innovation-en","tag-perplexity-en"],"_links":{"self":[{"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/posts\/32689","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/comments?post=32689"}],"version-history":[{"count":1,"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/posts\/32689\/revisions"}],"predecessor-version":[{"id":32690,"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/posts\/32689\/revisions\/32690"}],"wp:attachment":[{"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/media?parent=32689"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/categories?post=32689"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.mon-agent-ia.fr\/blog\/wp-json\/wp\/v2\/tags?post=32689"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}