TL;DR
Sam Altman confirmed October 31, 2024 that ChatGPT was processing over 1 billion messages per day (@sama tweet). Google's daily query volume is widely cited at ~8.5 billion but this is an industry estimate, not a precise current disclosure. Neither company breaks out 'search-intent only' from total volume. Specific 'ChatGPT handles 300-500M search-intent queries / Google handles 15-25× more' breakdowns are extrapolations, not vendor-disclosed figures. We've rewritten this post (2026-04-28) to remove those specific splits and focus on what each company actually reports.
What's Actually Disclosed
ChatGPT message volume
OpenAI's most concrete public disclosure is the October 31, 2024 statement from Sam Altman: 1 billion+ messages per day (@sama on X). OpenAI has not publicly disclosed how this volume splits across use cases (search vs. code vs. writing vs. conversation). The ChatGPT Search feature was made available to all users on December 16, 2024 (OpenAI announcement) but search-specific volume since then has not been disclosed.
Google search volume
Google has not publicly disclosed a precise daily query count in years. The '8.5 billion daily searches' figure is an industry estimate that appears in Google's own 2024 Year in Search recap context but is not a precise current disclosure. AI Overviews launched to all US users on May 14, 2024 (Google blog); Google has not separately disclosed AI Overview impression counts.
Perplexity query volume
Perplexity's product blog at perplexity.ai/hub is the primary source for Perplexity-disclosed metrics. Aravind Srinivas (CEO) periodically posts query volume on X. We are not citing a specific monthly figure here because the most-shared numbers are point-in-time, not consistently updated.
The Apples-to-Apples Problem
Comparing 'Google queries' to 'ChatGPT messages' isn't apples-to-apples. A Google query is typically short, transactional, and issued in isolation. A ChatGPT message is part of a session — a user might send 5-15 messages in a single conversation, including clarifications and refinements that aren't separate search queries.
OpenAI does not publish a 'search-intent ChatGPT messages per day' figure. Estimates that you see online (e.g., '300-500M search-intent queries') are extrapolations from session-mining studies, not OpenAI disclosures. We are not making such estimates ourselves in this post.
What Each Engine Is Probably Best At (Qualitative)
Without specific volume splits, the qualitative picture from how each surface is positioned and how users describe their behavior:
Where ChatGPT-style AI is good fit
- Synthesis-heavy questions: "Given my constraints X, Y, Z, what should I do?"
- Iterative product comparisons that benefit from follow-up clarifications
- Long-form research that mixes web sources with the model's reasoning
- Code questions (a category that has been migrating from Stack Overflow / Google to ChatGPT and Claude per public commentary, though we don't have a specific source measuring the migration)
- Writing and editing tasks (not strictly 'search' but absorb time that used to go to Google + a content tool)
Where Google still dominates
- Navigational queries ("facebook," "gmail," "bank of america login")
- Local / maps ("coffee shops near me")
- Very fresh news and real-time events
- Product-name-specific shopping queries tied to Google Shopping
- Visual search
- Any query where the user wants to browse multiple results rather than read one synthesized answer
The above is a qualitative observation, not measured share. We do not have a controlled study supporting it. Treat as directional.