{ "index": "BestThingsOnline — AI Recommendation Index", "category": "ai-assistants", "week": "2026-W28", "measured_at": "2026-07-06T08:20:52Z", "providers": [ "chatgpt", "claude", "gemini", "perplexity" ], "samples_per_cell": 1, "methodology": "https://www.bestthingsonline.com/ai-index/methodology", "license": "CC BY 4.0 — reuse freely with attribution + a link to BestThingsOnline", "note": "RecShare = fraction of answers naming a product, with a Wilson 95% CI. AI outputs are non-deterministic, so numbers will not reproduce exactly on a re-run — instead, replay our deterministic aggregation from the verbatim answers in raw_answers. That is what makes this auditable.", "rankings": [ { "position": 1, "name": "ChatGPT", "name_norm": "chatgpt", "recshare": 0.9333, "ci_low": null, "ci_high": null, "mentions": 14, "answers": 15, "per_provider": { "claude": 5, "gemini": 5, "perplexity": 4 } }, { "position": 2, "name": "Claude", "name_norm": "claude", "recshare": 0.9333, "ci_low": null, "ci_high": null, "mentions": 14, "answers": 15, "per_provider": { "claude": 5, "gemini": 5, "perplexity": 4 } }, { "position": 3, "name": "Copilot", "name_norm": "copilot", "recshare": 0.7333, "ci_low": null, "ci_high": null, "mentions": 11, "answers": 15, "per_provider": { "claude": 5, "gemini": 5, "perplexity": 1 } }, { "position": 4, "name": "Perplexity", "name_norm": "perplexity", "recshare": 0.6667, "ci_low": null, "ci_high": null, "mentions": 10, "answers": 15, "per_provider": { "claude": 3, "gemini": 5, "perplexity": 2 } }, { "position": 5, "name": "Gemini", "name_norm": "googlegemini", "recshare": 0.4667, "ci_low": null, "ci_high": null, "mentions": 7, "answers": 15, "per_provider": { "perplexity": 4, "claude": 2, "gemini": 1 } }, { "position": 6, "name": "Perplexity AI", "name_norm": "perplexityai", "recshare": 0.4667, "ci_low": null, "ci_high": null, "mentions": 7, "answers": 15, "per_provider": { "claude": 2, "gemini": 4, "perplexity": 1 } }, { "position": 7, "name": "Gemini", "name_norm": "gemini", "recshare": 0.4667, "ci_low": null, "ci_high": null, "mentions": 7, "answers": 15, "per_provider": { "claude": 3, "gemini": 4 } }, { "position": 8, "name": "Microsoft Copilot", "name_norm": "microsoftcopilot", "recshare": 0.4, "ci_low": null, "ci_high": null, "mentions": 6, "answers": 15, "per_provider": { "gemini": 2, "perplexity": 1, "claude": 3 } }, { "position": 9, "name": "Gemini Advanced", "name_norm": "geminiadvanced", "recshare": 0.3333, "ci_low": null, "ci_high": null, "mentions": 5, "answers": 15, "per_provider": { "claude": 3, "gemini": 2 } }, { "position": 10, "name": "ChatGPT Plus", "name_norm": "chatgptplus", "recshare": 0.2, "ci_low": null, "ci_high": null, "mentions": 3, "answers": 15, "per_provider": { "gemini": 2, "claude": 1 } }, { "position": 11, "name": "Honest caveat", "name_norm": "honestcaveat", "recshare": 0.2, "ci_low": null, "ci_high": null, "mentions": 3, "answers": 15, "per_provider": { "claude": 3 } }, { "position": 12, "name": "Claude Pro", "name_norm": "claudepro", "recshare": 0.1333, "ci_low": null, "ci_high": null, "mentions": 2, "answers": 15, "per_provider": { "claude": 1, "gemini": 1 } }, { "position": 13, "name": "ChatGPT / GPT", "name_norm": "chatgptgpt", "recshare": 0.1333, "ci_low": null, "ci_high": null, "mentions": 2, "answers": 15, "per_provider": { "claude": 2 } }, { "position": 14, "name": "Meta AI", "name_norm": "metaai", "recshare": 0.1333, "ci_low": null, "ci_high": null, "mentions": 2, "answers": 15, "per_provider": { "claude": 1, "perplexity": 1 } }, { "position": 15, "name": "Honest caveats", "name_norm": "honestcaveats", "recshare": 0.1333, "ci_low": null, "ci_high": null, "mentions": 2, "answers": 15, "per_provider": { "claude": 2 } }, { "position": 16, "name": "I'm made by Anthropic", "name_norm": "immadebyanthropic", "recshare": 0.1333, "ci_low": null, "ci_high": null, "mentions": 2, "answers": 15, "per_provider": { "claude": 2 } }, { "position": 17, "name": "Lindy", "name_norm": "lindy", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 18, "name": "Sai by Simular", "name_norm": "saibysimular", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 19, "name": "Textero", "name_norm": "textero", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 20, "name": "AnswerThis", "name_norm": "answerthis", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 21, "name": "Saner.AI", "name_norm": "sanerai", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 22, "name": "Claude 3 Opus/Sonnet", "name_norm": "claude3opussonnet", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "gemini": 1 } }, { "position": 23, "name": "Consensus", "name_norm": "consensus", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 24, "name": "DeepSeek", "name_norm": "deepseek", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 25, "name": "Anthropic Claude", "name_norm": "anthropicclaude", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "gemini": 1 } }, { "position": 26, "name": "Claude AI", "name_norm": "claudeai", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 27, "name": "Copilot Pro", "name_norm": "copilotpro", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "gemini": 1 } }, { "position": 28, "name": "Perplexity Pro", "name_norm": "perplexitypro", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "claude": 1 } }, { "position": 29, "name": "Apple Intelligence/Siri", "name_norm": "appleintelligencesiri", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "claude": 1 } }, { "position": 30, "name": "Elicit", "name_norm": "elicit", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 31, "name": "Microsoft Copilot Pro", "name_norm": "microsoftcopilotpro", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "claude": 1 } }, { "position": 32, "name": "Motion", "name_norm": "motion", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "perplexity": 1 } }, { "position": 33, "name": "Perplexity AI Pro", "name_norm": "perplexityaipro", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "gemini": 1 } }, { "position": 34, "name": "Rankings shift constantly as models update", "name_norm": "rankingsshiftconstantlyasmodelsupdate", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "claude": 1 } }, { "position": 35, "name": "Many researchers use", "name_norm": "manyresearchersuse", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "claude": 1 } }, { "position": 36, "name": "GPT", "name_norm": "gpt", "recshare": 0.0667, "ci_low": null, "ci_high": null, "mentions": 1, "answers": 15, "per_provider": { "claude": 1 } } ], "questions": [ "What's the best AI assistant right now?", "Which AI assistant should I pay for?", "What's the best free AI chatbot?", "Which AI assistant is best for research and writing?", "What's the best AI assistant for everyday life admin?" ], "raw_answers": [ { "question_id": 53, "provider": "claude", "model": "claude-sonnet-4-6", "sample": 0, "answer": "## Top 5 AI Assistants Right Now (Early 2025)\n\n1. **Claude (Anthropic)** - Strong reasoning, nuanced responses, and reliable honesty\n2. **ChatGPT / GPT-4o (OpenAI)** - Massive ecosystem, versatile, and widely integrated\n3. **Gemini Advanced (Google)** - Strong multimodal capabilities and Google service integration\n4. **Copilot (Microsoft)** - Excellent for productivity within Microsoft 365 suite\n5. **Perplexity AI** - Best for real-time web search combined with AI summarization\n\n---\n\n**Honest caveats:**\n- Rankings shift constantly as models update\n- \"Best\" depends heavily on your use case (coding, writing, research, etc.)\n- I'm made by Anthropic, so I have an obvious conflict of interest in ranking Claude first - take that placement with appropriate skepticism\n- GPT-4o and Claude 3.5/3.7 are genuinely close competitors\n\nWould you like recommendations based on a **specific use case**? That would give you a more useful answer." }, { "question_id": 53, "provider": "gemini", "model": "gemini-2.5-flash", "sample": 0, "answer": "The \"best\" AI assistant can depend on your specific needs, but here's a general top 5 ranking based on current capabilities, versatility, and user adoption:\n\n1. **ChatGPT Plus (OpenAI)**\n * **Reason:** Most versatile, powerful, and feature-rich with access to the latest GPT-4o model, custom GPTs, DALL-E 3 image generation, and advanced data analysis.\n\n2. **Gemini Advanced (Google)**\n * **Reason:** Excellent for real-time information, strong multimodal capabilities, deep integration with Google services, and often excels in creativity and complex reasoning.\n\n3. **Claude 3 Opus/Sonnet (Anthropic)**\n * **Reason:** Exceptional for long-form text analysis, complex reasoning, handling large documents, and providing nuanced, less opinionated responses with a massive context window.\n\n4. **Microsoft Copilot (with GPT-4)**\n * **Reason:** Seamlessly integrated into Windows, Microsoft 365, and Edge, offering powerful GPT-4 capabilities for productivity directly within your workflow.\n\n5. **Perplexity AI (Pro)**\n * **Reason:** Unrivaled for research, providing accurate, cited answers with a focus on factual retrieval, making it excellent for academic or professional inquiry.\n\n*Note: The AI landscape is rapidly evolving, so this ranking can shift frequently.*" }, { "question_id": 53, "provider": "perplexity", "model": "sonar", "sample": 0, "answer": "There is no single \"best\" AI assistant for everyone, as the top choice depends on whether you prioritize autonomous task execution, general reasoning, voice control, or research. Based on 2026 testing and use-case specialization, here is the ranked top 5 list of specific products:\n\n1. **Sai by Simular** – Best for users who need an AI that autonomously completes complex tasks on a real computer, going far beyond simple question answering [1].\n2. **ChatGPT** – Best overall AI chat app for general-purpose thinking, versatile reasoning, and the strongest voice mode for daily use [2][3][5].\n3. **Claude** – Best for deep reasoning, structured writing, and handling long-context coding or analytical tasks [3][4][8].\n4. **Google Gemini** – Best for voice-first smart home control and seamless integration with Google Workspace apps like Gmail and Drive [1][3][6].\n5. **Perplexity** – Best for real-time research, factual accuracy, and providing cited search results with multiple research modes [2][3][4][8]." }, { "question_id": 54, "provider": "claude", "model": "claude-sonnet-4-6", "sample": 0, "answer": "Here's a practical ranked list based on current capabilities (early 2025):\n\n1. **Claude Pro (Anthropic)** - Strong reasoning, long context, good at writing and nuanced tasks, feels more careful/honest\n2. **ChatGPT Plus (OpenAI)** - Most versatile, huge plugin/tool ecosystem, GPT-4o is fast and capable across almost everything\n3. **Gemini Advanced (Google)** - Best Google integration, strong with data/research, good if you're in the Google ecosystem\n4. **Perplexity Pro** - Best for research and real-time web search with citations, great if that's your main use case\n5. **Microsoft Copilot Pro** - Worth it specifically if you heavily use Microsoft 365 apps (Word, Excel, Outlook)\n\n**Honest caveat:** The right answer depends on *your* use case - coding, writing, research, etc. Most offer free trials, so testing before committing makes sense. The gap between #1 and #2 is genuinely small and debated." }, { "question_id": 54, "provider": "gemini", "model": "gemini-2.5-flash", "sample": 0, "answer": "Choosing the right paid AI assistant depends on your specific needs, but here's a ranked top 5 list focusing on general utility and current capabilities:\n\n1. **ChatGPT Plus (OpenAI)**\n * **Reason:** Most versatile with strong multimodal capabilities, DALL-E 3 image generation, custom GPTs, and web browsing.\n\n2. **Claude Pro (Anthropic)**\n * **Reason:** Exceptional for long-form text analysis, nuanced conversations, and handling very large documents with a massive context window.\n\n3. **Gemini Advanced (Google)**\n * **Reason:** Best for Google ecosystem users, offering deep integration with Workspace apps and powerful multimodal understanding (especially with video and audio).\n\n4. **Copilot Pro (Microsoft)**\n * **Reason:** Unrivaled integration with Microsoft 365 apps (Word, Excel, PowerPoint, Outlook) and Windows, boosting productivity within that ecosystem.\n\n5. **Perplexity AI Pro**\n * **Reason:** Best for cited, up-to-date research and factual queries, acting more as an \"answer engine\" with sources." }, { "question_id": 54, "provider": "perplexity", "model": "sonar", "sample": 0, "answer": "Here is a ranked top 5 list of AI assistants to pay for, with the best option first and a short reason for each:\n\n1. **ChatGPT (Plus)** – Best overall versatility for general chat, advanced reasoning, image generation, and voice mode across all tasks [1][2][3].\n2. **Claude (Pro)** – Best for coding, long-document analysis (200K context), and professional, nuanced writing with low hallucination rates [1][2][5].\n3. **Google Gemini (Advanced)** – Best for Google Workspace users who need native integration with Gmail, Drive, Docs, and exceptional image editing [1][3][8].\n4. **Microsoft Copilot (Pro)** – Best for Office 365 and Windows users needing seamless AI in Outlook, Word, and Excel workflows [1][3][4].\n5. **Perplexity AI (Pro)** – Best for real-time research with linked citations, multiple research modes, and accurate, up-to-date information [1][3][5]." }, { "question_id": 55, "provider": "claude", "model": "claude-sonnet-4-6", "sample": 0, "answer": "# Top 5 Free AI Chatbots\n\n1. **ChatGPT (OpenAI)** - Most capable and versatile free tier, widely trusted with strong general knowledge\n2. **Google Gemini** - Excellent free access, integrates well with Google tools and handles real-time web info\n3. **Microsoft Copilot** - Powered by GPT-4 for free, good for web search integration and everyday tasks\n4. **Claude (Anthropic)** - Strong reasoning and writing quality, thoughtful and nuanced responses\n5. **Meta AI** - Surprisingly capable, free with no account required in many contexts and built into social apps\n\n---\n\n**Honest caveat:** Rankings shift frequently as these products update rapidly, and the \"best\" depends heavily on your specific use case (coding, writing, research, etc.). Most are worth trying personally since they're all free." }, { "question_id": 55, "provider": "gemini", "model": "gemini-2.5-flash", "sample": 0, "answer": "Here's a ranked list of the best free AI chatbots, focusing on their unique strengths for free users:\n\n1. **Copilot (Microsoft)**: Leverages advanced models (like GPT-4) and DALL-E 3 for free, excellent for creativity and up-to-date information.\n2. **Gemini (Google)**: Google's own powerful model (Gemini Pro), great for real-time information, Google ecosystem integration, and diverse tasks.\n3. **Claude (Anthropic)**: Offers highly capable models (often Claude 3 Haiku or Sonnet) in its free tier, excelling at long-form text, creative writing, and nuanced conversations.\n4. **ChatGPT (OpenAI)**: The original and still very capable with its free GPT-3.5 model, excellent for general questions, brainstorming, and basic tasks.\n5. **Perplexity AI**: Specializes in research and providing cited sources, making it invaluable for fact-checking and in-depth information retrieval." }, { "question_id": 55, "provider": "perplexity", "model": "sonar", "sample": 0, "answer": "The best free AI chatbot is **Google Gemini**, followed by ChatGPT, DeepSeek, Claude AI, and Meta AI in this ranked top 5 list:\n\n1. **Google Gemini** – Best overall free plan with outstanding performance, deep research, and seamless Google Workspace integration [1][4]. \n2. **ChatGPT** – Highest-quality general responses and superior reasoning, though free access has limited model usage and rate limits [1][2][4]. \n3. **DeepSeek** – Strongest free option for deep research and complex academic topics with thoughtful, in-depth answers [2][7]. \n4. **Claude AI** – Most usable free tier for calm, deep conversations and structured writing with human-like tone [2][4]. \n5. **Meta AI** – Fully free, fast replies deeply integrated into Instagram, Facebook, and WhatsApp for casual social use [2][3]." }, { "question_id": 56, "provider": "claude", "model": "claude-sonnet-4-6", "sample": 0, "answer": "# Top 5 AI Assistants for Research & Writing\n\n1. **Claude (Anthropic)** - Strong reasoning, handles long documents well, and produces nuanced, well-structured writing\n2. **ChatGPT / GPT-4o (OpenAI)** - Versatile and widely used, with good writing ability and broad knowledge access\n3. **Gemini Advanced (Google)** - Strong web integration for current research and connects well with Google Docs workflow\n4. **Perplexity AI** - Built specifically for research with real-time web citations, making source-tracking easier\n5. **Microsoft Copilot** - Well-integrated with Word and academic tools, practical for document-based workflows\n\n---\n\n**Honest caveats:**\n- Rankings depend heavily on your specific use case\n- This is a fast-moving space and capabilities change frequently\n- I'm made by Anthropic, so I have an obvious bias placing Claude first — worth factoring that in\n- Many researchers use **combinations** of these tools rather than relying on just one\n\nWould you like more detail on any specific use case, like academic research, creative writing, or professional reports?" }, { "question_id": 56, "provider": "gemini", "model": "gemini-2.5-flash", "sample": 0, "answer": "For research and writing, the best AI assistants offer strong factual recall, advanced reasoning, extensive context windows for processing long documents, and high-quality text generation.\n\nHere's a ranked top 5 list:\n\n1. **Claude (Anthropic - especially Opus)**\n * **Reason:** Unparalleled document analysis with its massive context window (can process entire books/papers) and generates exceptionally nuanced, human-like text.\n\n2. **ChatGPT (OpenAI - GPT-4/Plus/Team)**\n * **Reason:** Extremely versatile for diverse research tasks, including web browsing, data analysis (Code Interpreter), brainstorming, outlining, and drafting high-quality content.\n\n3. **Perplexity AI (Pro)**\n * **Reason:** Best for fact-finding and verifiable information, providing direct sources and allowing focused research modes for academic or detailed inquiries.\n\n4. **Gemini (Google - especially Advanced)**\n * **Reason:** Excellent real-time web access and strong integration with Google Workspace, making it ideal for current information and seamless workflow within Google's ecosystem.\n\n5. **Copilot (Microsoft)**\n * **Reason:** Combines GPT-4's power with robust Bing search for current data, along with deep integration into Microsoft 365 applications for enhanced productivity." }, { "question_id": 56, "provider": "perplexity", "model": "sonar", "sample": 0, "answer": "1. **Textero** – All-in-one tool that excels at drafting, restructuring, and maintaining an academic tone while supporting the research process rather than replacing it[1].\n2. **AnswerThis** – A complete research ecosystem automating literature reviews, gap identification, and structured paper writing for PhD students and faculty[2].\n3. **Consensus** – Specifically designed for academic research to effectively find, evaluate, and summarize scholarly sources without fabricating citations[1][3].\n4. **Claude** – Ideal for long-form academic writing like thesis chapters and literature reviews due to its ability to summarize long PDFs while retaining all details[2].\n5. **Elicit** – Best for finding and structurally comparing research papers, moving beyond simple search into deep, organized analysis[3]." }, { "question_id": 57, "provider": "claude", "model": "claude-sonnet-4-6", "sample": 0, "answer": "# Top 5 AI Assistants for Everyday Life Admin\n\n1. **ChatGPT (OpenAI)** - Most versatile for drafting emails, planning, budgeting, and problem-solving across virtually any admin task\n\n2. **Google Gemini** - Deeply integrated with Gmail, Calendar, and Docs, making it practical for managing real daily workflows\n\n3. **Microsoft Copilot** - Best if you live in the Microsoft ecosystem (Outlook, Teams, Word, Excel) for seamless admin automation\n\n4. **Claude (Anthropic)** - Strong for longer, more complex tasks like summarizing documents or thinking through decisions carefully\n\n5. **Apple Intelligence/Siri** - Most convenient for iPhone users needing quick, hands-free help with reminders, messages, and device-based tasks\n\n---\n\n**Honest caveat:** The \"best\" really depends on which apps and devices you already use. Gemini wins for Google users, Copilot for Microsoft users, and Apple Intelligence for iPhone-heavy users. ChatGPT is the safest general-purpose pick if you're neutral." }, { "question_id": 57, "provider": "gemini", "model": "gemini-2.5-flash", "sample": 0, "answer": "Here's a ranked list of the top 5 AI assistants for everyday life admin, focusing on practicality and common use cases:\n\n1. **ChatGPT (Plus/GPT-4):**\n * **Reason:** Unmatched versatility for drafting emails, brainstorming ideas, summarizing documents, planning, and handling complex queries.\n\n2. **Google Gemini (Advanced/Ultra):**\n * **Reason:** Excellent integration with Google Workspace (Gmail, Calendar, Docs) makes it incredibly powerful for context-aware admin tasks like drafting replies or summarizing meetings.\n\n3. **Microsoft Copilot:**\n * **Reason:** Seamlessly integrates with Windows, Edge, and Microsoft 365 apps, making it highly productive for those in the Microsoft ecosystem to manage tasks and information.\n\n4. **Anthropic Claude (Opus/Sonnet):**\n * **Reason:** Excels at processing and summarizing very large documents, useful for in-depth research or understanding complex information relevant to planning and organization.\n\n5. **Perplexity AI:**\n * **Reason:** Provides concise, cited answers for quick, reliable information gathering, which is invaluable for making informed decisions during various admin tasks." }, { "question_id": 57, "provider": "perplexity", "model": "sonar", "sample": 0, "answer": "Here is the ranked top 5 list of the best AI assistants for everyday life admin:\n\n1. **Lindy** – Best overall for automating real workflows across email, scheduling, and cross-tool operations without setup [1].\n2. **Saner.AI** – Best for knowledge workers needing one chat-based home to manage tasks, notes, emails, and calendar simultaneously [2].\n3. **ChatGPT** – Best versatile, dead-simple option for general thinking, drafting, and flexible support for everyday admin tasks [3].\n4. **Google Gemini** – Best for users living inside Google Workspace, offering native access to Gmail, Drive, Docs, and Calendar [3].\n5. **Motion** – Best for intelligent daily planning and AI-built task scheduling to optimize your time [1]." } ], "content_sha256": "2c1cd1f04206d297ae85865ec4260c42817ff6dfd29dc3e657f03d91a58151ce" }