We measure how belief is made: in a story, in an idea, in a live event.

AI-SCOPE · A DOXASCOPE LENS

What does the free plan give you, and what are its limits?

What you get today without paying, service by service: model, limit, tools, account rules. Each value shows how we know it and when it was last verified.

9 services118 findingsFindings checked: to Record assembled:
ChatGPT17 findings

  • Default model

    Confirmed

    OpenAI's help article on GPT-5.6 says the default model for Free and Go users is GPT-5.6 Luna, which also powers Think. The release notes of 14 August 2026 announced the change for that week.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-DEFAULT-MODEL

  • Available models

    Confirmed

    ChatGPT's plan-comparison table marks GPT-5.6 Luna and GPT-5 Thinking Mini Yes for the Free plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-AVAILABLE-MODELS-2

  • Available models

    Confirmed

    ChatGPT's plan-comparison table gives the Free plan GPT-5.6 Terra only as limited access, in Work and Codex on desktop.

    Interface:
    Desktop app
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-AVAILABLE-MODELS-3

  • Usage limit

    Confirmed

    Everyday text chats on ChatGPT Free are unlimited in count, subject only to abuse-prevention guardrails.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-USAGE-LIMIT

  • Usage limit

    Confirmed

    ChatGPT Free includes 500 MB of storage in the Library feature.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-USAGE-LIMIT-3

  • Usage limit

    Confirmed

    On ChatGPT Free, image generation, voice, data analysis and other tools have usage limits separate from the main chat limit, and OpenAI publishes no figures for them. File uploads are the exception, with a published limit of 3 a day.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-USAGE-LIMIT-4

  • Usage limit

    Confirmed

    OpenAI's File Uploads FAQ says Free users are limited to 3 file uploads per day, and that OpenAI may lower the limits during peak hours.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-USAGE-LIMIT-5

  • Image generation

    Confirmed

    ChatGPT Free tier users can create images in ChatGPT, though limited and slower than paid plans.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Confirmed

    ChatGPT Free tier users can upload images or files in their prompts, subject to a separate limit.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-TOOL-FILE-UPLOAD

  • Voice

    Confirmed

    Voice is available to ChatGPT Free tier users, though with a limited number of voice chats.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-TOOL-VOICE

  • Context window

    Confirmed

    The total context window for the Instant model on ChatGPT Free is 27K tokens, per OpenAI's pricing page.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-CONTEXT-WINDOW

  • Context window

    Not disclosed

    For the Reasoning model, OpenAI publishes no context-window figure for the Free plan; the pricing page marks it "Varies" instead of a number.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-FREE-CONTEXT-WINDOW-2

  • Minimum age

    Confirmed

    OpenAI's Terms of Use require users to be at least 13 years old, or the minimum age required in their country, with parental permission required under 18.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-ALL-ACCOUNT-MIN-AGE

  • Phone requirement

    Confirmed

    OpenAI's help center says phone verification is no longer required to create a new OpenAI account or to use ChatGPT. It is still required when generating a first API key.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-CHATGPT-ALL-ACCOUNT-PHONE-2

  • Country availability

    Confirmed

    OpenAI's official Supported Countries page names 21 of the 22 Arab League countries among the countries and territories where it currently supports access to ChatGPT.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Supported
    • Bahrain: Supported
    • Djibouti: Supported
    • Algeria: Supported
    • Egypt: Supported
    • Iraq: Supported
    • Jordan: Supported
    • Comoros: Supported
    • Kuwait: Supported
    • Lebanon: Supported
    • Libya: Supported
    • Morocco: Supported
    • Mauritania: Supported
    • Oman: Supported
    • Palestine: Supported
    • Qatar: Supported
    • Saudi Arabia: Supported
    • Sudan: Supported
    • Somalia: Supported
    • Syria: Not established by this finding
    • Tunisia: Supported
    • Yemen: Supported
    F-CHATGPT-ALL-COUNTRY-AVAILABILITY-1

  • Country availability

    Confirmed

    Syria does not appear on OpenAI's official list of supported countries and territories, and the page warns that access from outside the listed countries may get an account blocked.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Not established by this finding
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Not established by this finding
    • Egypt: Not established by this finding
    • Iraq: Not established by this finding
    • Jordan: Not established by this finding
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Not established by this finding
    • Morocco: Not established by this finding
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Not established by this finding
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not supported
    • Tunisia: Not established by this finding
    • Yemen: Not established by this finding
    F-CHATGPT-ALL-COUNTRY-AVAILABILITY-2
Gemini13 findings

  • Default model

    Unknown

    Google's Gemini Apps help pages list three models available without an AI plan (Flash-Lite, Flash and Pro) but do not state which one is used by default for a new chat.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • Gemini Apps limits & upgrades for Google AI subscribers (support.google.com/gemini/answer/16275805) — lists three available models for 'Without an AI Plan' with no stated default; its Feature-availability table (which does distinguish Yes/No per plan via alt='Checkmark' vs alt='') has no default-model row
    • Google AI plans page (one.google.com/about/google-ai-plans) — describes paid-plan model access tiers, no default-model statement for users without a plan
    • What you need to sign in to Gemini Apps (support.google.com/gemini/answer/13278668) — no mention of a default model
    • Gemini Apps release updates (gemini.google.com/updates), read 2026-09-22 in browser mode (served in Arabic) — release notes back to 2023, no statement of the default model for users without a plan
    • Gemini app listings on the App Store and Google Play (captured 2026-09-22) — name "our latest models", not a default for the free plan
    F-GEMINI-FREE-DEFAULT-MODEL

  • Available models

    Confirmed

    Gemini's Model Access table marks Gemini 3 Flash-lite, Gemini 3 Flash and Gemini 3 Pro as available (Yes) to users without a Google AI plan, the same as every paid plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-FREE-AVAILABLE-MODELS

  • Usage limit

    Confirmed

    Gemini Apps usage limits are compute-based; Google states the limit refreshes every 5 hours until a weekly cap is reached, without publishing the underlying figures for users without a Google AI plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-FREE-USAGE-LIMIT

  • Web search

    Confirmed

    Gemini's feature-availability table marks Deep Research with a checkmark in the 'Without a Google AI plan' column, and Google's Deep Research help articles (computer, Android and iPhone) say Gemini includes Google Search as a source by default. The evidence covers web search within Deep Research only; no captured page names a web-search tool for ordinary chat. Google also says Deep Research may be unavailable to users without a plan during periods of high demand.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-FREE-TOOL-WEB-SEARCH-2

  • Image generation

    Confirmed

    Gemini's feature-availability table marks 'Image generation with Nano Banana 2' with a checkmark in the 'Without a Google AI plan' column, meaning the feature is available for accounts without a Google AI plan; other rows in the same table (such as 'Video generation' and 'Redo images with Nano Banana Pro') leave that column empty, showing the table does distinguish yes from no.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Confirmed

    Gemini's help page describes uploading files to Gemini Apps without tying the feature to a paid plan, and Google publishes a context-window figure (32k tokens) specifically for users without a Google AI plan for that purpose, confirming the feature is available to them.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-FREE-TOOL-FILE-UPLOAD

  • Voice

    Confirmed

    Gemini's Google Play listing names Gemini Live among the app's features without tying it to a paid plan, while it ties video creation to Google AI Plus, Pro and Ultra. Google's help article lists what Gemini Live needs (an Android device, the app, a Google account, age 13 or over) with no paid plan among them.

    Interface:
    Android
    Plan:
    Free
    How to read this lens
    Observed official product interface
    Last checked:
    F-GEMINI-FREE-TOOL-VOICE-2

  • Voice

    Confirmed

    Gemini's App Store listing names Gemini Live among the app's features without tying it to a paid plan, while it ties video creation to Google AI Plus, Pro and Ultra. Google's help article lists what Gemini Live needs (an iPhone or iPad, the app, a Google account, age 13 or over) with no paid plan among them.

    Interface:
    iOS
    Plan:
    Free
    How to read this lens
    Observed official product interface
    Last checked:
    F-GEMINI-FREE-TOOL-VOICE-3

  • Context window

    Confirmed

    For users without a Google AI plan, Gemini Apps' context window is 32k tokens, per Google's help page.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-FREE-CONTEXT-WINDOW

  • Minimum age

    Confirmed

    To sign in to Gemini Apps with a personal or school Google Account, Google requires users to be 13 (or the applicable minimum age in their country) or older.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-ALL-ACCOUNT-MIN-AGE

  • Phone requirement

    Confirmed

    Gemini needs a Google Account. Google's help on creating one lists adding a phone number as an optional step, and allows signing up with an existing non-Google email address.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-GEMINI-ALL-ACCOUNT-PHONE-2

  • Country availability

    Confirmed

    Google's 'Where you can use the Gemini web app' help page, under the heading 'Supported countries & territories', names 21 of the 22 Arab League countries.

    Interface:
    Web
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Supported
    • Bahrain: Supported
    • Djibouti: Supported
    • Algeria: Supported
    • Egypt: Supported
    • Iraq: Supported
    • Jordan: Supported
    • Comoros: Supported
    • Kuwait: Supported
    • Lebanon: Supported
    • Libya: Supported
    • Morocco: Supported
    • Mauritania: Supported
    • Oman: Supported
    • Palestine: Supported
    • Qatar: Supported
    • Saudi Arabia: Supported
    • Sudan: Supported
    • Somalia: Supported
    • Syria: Not established by this finding
    • Tunisia: Supported
    • Yemen: Supported
    F-GEMINI-ALL-COUNTRY-AVAILABILITY-1

  • Country availability

    Confirmed

    Syria does not appear in Google's 'Supported countries & territories' list for the Gemini web app; since the page presents itself as a complete list under that heading, this fact reads the omission as unsupported rather than merely unstated.

    Interface:
    Web
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Not established by this finding
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Not established by this finding
    • Egypt: Not established by this finding
    • Iraq: Not established by this finding
    • Jordan: Not established by this finding
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Not established by this finding
    • Morocco: Not established by this finding
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Not established by this finding
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not supported
    • Tunisia: Not established by this finding
    • Yemen: Not established by this finding
    F-GEMINI-ALL-COUNTRY-AVAILABILITY-2
Meta AI13 findings

  • Default model

    Confirmed

    Meta AI runs on Muse Spark from Meta Superintelligence Labs by default, per the FAQ on Meta's official Meta AI page.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-META-AI-FREE-DEFAULT-MODEL

  • Available models

    Confirmed

    Meta's official Meta AI page names one model, Muse Spark, and offers Instant and Thinking as response modes of that same model, not as separate models.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-META-AI-FREE-AVAILABLE-MODELS

  • Usage limit

    Confirmed

    Meta says it has started testing usage limits on some of Meta AI's more compute-intensive features, on an unnumbered daily-or-monthly cadence; on hitting the limit, users are offered a subscription to unlock more usage before the free limit resets.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-META-AI-FREE-USAGE-LIMIT

  • Image generation

    Confirmed

    Meta AI users can generate images, per the FAQ on Meta's official Meta AI page.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-META-AI-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Confirmed

    Meta AI users can upload files and images, such as a lease, a spreadsheet, or product photos to compare, per Meta's official Meta AI page.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-META-AI-FREE-TOOL-FILE-UPLOAD

  • Voice

    Confirmed

    Meta AI offers a conversational voice mode that supports 13 languages, per Meta's official Meta AI page.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-META-AI-FREE-TOOL-VOICE

  • Context window

    Unknown

    Meta publishes no context-window or token-limit figure for Meta AI on any of the official pages reviewed.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • Meta AI official marketing/FAQ page (ai.meta.com/meta-ai/) — describes model, tools and modes, states no context-window or token-limit figure
    • Meta Help Center article on Meta AI account use (meta.com/help/artificial-intelligence/31983416211257755/) — covers phone/email use, not context window
    • Web search: "Meta AI context window tokens limit official help.meta.com" — returned only unrelated Meta Help Center pages (privacy requests, harm reporting) and third-party context-window explainers, no Meta-owned page on Meta AI's context window
    • WhatsApp Help Center, 'About Meta AI' (faq.whatsapp.com/2257017191175152), read 2026-09-22 — describes features, no context-window figure
    • whatsapp.com/meta-ai, read 2026-09-22 — no context-window figure
    F-META-AI-FREE-CONTEXT-WINDOW

  • Minimum age

    Confirmed

    Under Meta AI's Terms of Service, access to Meta AI is not permitted for anyone under age 13, or the greater age required in their country or territory.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-META-AI-ALL-ACCOUNT-MIN-AGE

  • Phone requirement

    Confirmed

    Meta says Meta AI needs a Meta Account, and that Meta Accounts work on Facebook and Meta AI. Facebook's sign-up help asks for an email address or a mobile number, and confirms either one. This does not cover Meta AI inside WhatsApp.

    Interface:
    Web
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-META-AI-ALL-ACCOUNT-PHONE-2

  • Country availability

    Confirmed

    Two press reports (TechCrunch and Wamda) covered Meta AI's Arabic launch in February 2025; both name the UAE, Saudi Arabia, Egypt, Morocco and Iraq. Meta publishes no official country list.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Independent source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Supported
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Not established by this finding
    • Egypt: Supported
    • Iraq: Supported
    • Jordan: Not established by this finding
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Not established by this finding
    • Morocco: Supported
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Supported
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not established by this finding
    • Tunisia: Not established by this finding
    • Yemen: Not established by this finding
    F-META-AI-ALL-COUNTRY-AVAILABILITY-3

  • Country availability

    Single source

    TechCrunch's report (24 February 2025) alone names Algeria, Jordan, Libya, Tunisia and Yemen among the countries where Meta AI would be available. The second report and Meta's own pages do not name them.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Independent source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Not established by this finding
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Supported
    • Egypt: Not established by this finding
    • Iraq: Not established by this finding
    • Jordan: Supported
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Supported
    • Morocco: Not established by this finding
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Not established by this finding
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not established by this finding
    • Tunisia: Supported
    • Yemen: Supported
    F-META-AI-ALL-COUNTRY-AVAILABILITY-4

  • Country availability

    Not disclosed

    WhatsApp's help center says Meta AI is currently available only in "limited countries" and is rolling out gradually, without listing the countries.

    Interface:
    WhatsApp
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-META-AI-ALL-COUNTRY-AVAILABILITY-2
DeepSeek17 findings

  • Default model

    Unknown

    No available official source states which model powers the free chat app at chat.deepseek.com by default; the page itself could not be reviewed (see tried).

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • chat.deepseek.com, plain fetch — HTTP 202 with an empty (0-byte) body, no text to read
    • chat.deepseek.com, browser fetch — returns DeepSeek's bot "Human Verification" challenge page ("Before proceeding to your request, you need to solve a puzzle…"), not the consumer chat product (captures/2026-09-15/src-deepseek-03.html, kept as history but removed from sources.tsv per the controller's ruling)
    • chat.deepseek.com, Wayback Machine — resolved a snapshot (20260915065927) but the decoded text was 8 characters, too short to use
    • DeepSeek API docs (api-docs.deepseek.com/quick_start/pricing/) — lists API model names (deepseek-flash, deepseek-v4-pro) for API calls, but does not state which model the free consumer chat app uses by default
    • DeepSeek Terms of Use (cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html) — describes the models generically as "large language models" and does not name a default model
    • Web search "DeepSeek chat app default model FAQ help.deepseek.com" — returned no help.deepseek.com or other DeepSeek-owned page naming a default consumer-app model
    • DeepSeek change log (api-docs.deepseek.com/updates/), read 2026-09-22 — says V4-Pro GA "has been rolled out on the APP, Web, and API" (13 Aug 2026); the 10 Sep 2026 V4.1-Flash entry covers the API only; neither names a default for the chat app
    • DeepSeek App Store listing (captured 2026-09-22) — "Powered by DeepSeek’s latest flagship model", unnamed; version notes of 11–12 Sep 2026 say "Instant, Expert, and Vision modes are now unified"
    F-DEEPSEEK-FREE-DEFAULT-MODEL

  • Available models

    Unknown

    No current official source names the models available in DeepSeek's free app. The change log said on 13 August 2026 that DeepSeek-V4-Pro was rolled out on the app and web; app notes in September 2026 then said the Instant, Expert and Vision modes were "unified", without naming a model.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • chat.deepseek.com, plain, browser and Wayback captures (2026-09-15) — bot challenge or empty body; the chat app cannot be preserved (PROBE.md)
    • DeepSeek API pricing page (api-docs.deepseek.com/quick_start/pricing/, captured 2026-09-22) — covers API models and limits only, not the free chat app
    • DeepSeek change log (api-docs.deepseek.com/updates/), read 2026-09-22 — says V4-Pro GA "has been rolled out on the APP, Web, and API" (13 Aug 2026); the 10 Sep 2026 V4.1-Flash entry covers the API only
    • DeepSeek App Store listing (captured 2026-09-22) — "Powered by DeepSeek’s latest flagship model", unnamed; version notes of 11–12 Sep 2026 say "Instant, Expert, and Vision modes are now unified"
    F-DEEPSEEK-FREE-AVAILABLE-MODELS

  • Available models

    Confirmed

    This is from DeepSeek's API documentation, not a statement about the free chat app. The page lists two model names available through the API: deepseek-flash and deepseek-v4-pro.

    Interface:
    API
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-AVAILABLE-MODELS

  • Usage limit

    Unknown

    We found no official source stating a usage limit for DeepSeek's free app. The published limits (concurrent requests) apply to the API only.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • chat.deepseek.com, plain, browser and Wayback captures (2026-09-15) — bot challenge or empty body; the chat app cannot be preserved (PROBE.md)
    • DeepSeek API pricing page (api-docs.deepseek.com/quick_start/pricing/, captured 2026-09-22) — covers API models and limits only, not the free chat app
    • DeepSeek app launch notice (api-docs.deepseek.com/news/news250115/, captured 2026-09-22) — "100% FREE - No ads, no in-app purchases"; no usage limit stated
    • DeepSeek Terms of Use (cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html, captured 2026-09-22) — no usage limit for the free chat app
    F-DEEPSEEK-FREE-USAGE-LIMIT

  • Usage limit

    Confirmed

    This is from DeepSeek's API documentation, not a statement about the free chat app's limits. The pricing page states a concurrency limit of 2500 concurrent requests for the deepseek-flash API model, with no time window given.

    Interface:
    API
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-USAGE-LIMIT-1

  • Usage limit

    Confirmed

    This is from DeepSeek's API documentation, not a statement about the free chat app's limits. The pricing page states a concurrency limit of 500 concurrent requests for the deepseek-v4-pro API model, with no time window given.

    Interface:
    API
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-USAGE-LIMIT-2

  • Web search

    Confirmed

    DeepSeek's app launch notice (15 January 2025) lists web search among the features of the free app on the App Store and Google Play. We found no newer official page that restates it.

    Interface:
    Android
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-FREE-TOOL-WEB-SEARCH-3

  • Web search

    Confirmed

    DeepSeek's app launch notice (15 January 2025) lists web search among the features of the free app on the App Store and Google Play. We found no newer official page that restates it.

    Interface:
    iOS
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-FREE-TOOL-WEB-SEARCH-2

  • Image generation

    Unknown

    No available official source states whether the free chat app offers an image-generation tool.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • chat.deepseek.com, plain fetch — HTTP 202 with an empty (0-byte) body, no text to read
    • chat.deepseek.com, browser fetch — returns DeepSeek's bot "Human Verification" challenge page ("Before proceeding to your request, you need to solve a puzzle…"), not the consumer chat product (captures/2026-09-15/src-deepseek-03.html, kept as history but removed from sources.tsv per the controller's ruling)
    • chat.deepseek.com, Wayback Machine — resolved a snapshot (20260915065927) but the decoded text was 8 characters, too short to use
    • DeepSeek API docs (api-docs.deepseek.com/quick_start/pricing/) — lists "Vision" as a feature for one API model (image understanding as input), not image generation, and does not describe the free chat app
    • DeepSeek Terms of Use (cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html) — describes Outputs generically as "text, tables, and code"; does not mention image generation
    • Web search "DeepSeek chat app default model FAQ help.deepseek.com" — returned no help.deepseek.com or other DeepSeek-owned page describing an image-generation tool
    • DeepSeek app launch notice (api-docs.deepseek.com/news/news250115/, captured 2026-09-22) — lists web search, Deep-Think and file upload; no image generation
    • DeepSeek App Store listing (captured 2026-09-22) — version notes mention "vision mode" (image understanding), not image generation
    F-DEEPSEEK-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Confirmed

    DeepSeek's app launch notice (15 January 2025) lists file upload and text extraction among the features of the free app on the App Store and Google Play.

    Interface:
    Android
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-FREE-TOOL-FILE-UPLOAD-3

  • File upload

    Confirmed

    DeepSeek's App Store listing describes the app as free, and one of its version notes (July 2026) reads "Improved photo and file uploads". The app launch notice (January 2025) lists file upload and text extraction among its features.

    Interface:
    iOS
    Plan:
    Free
    How to read this lens
    Observed official product interface
    Last checked:
    F-DEEPSEEK-FREE-TOOL-FILE-UPLOAD-2

  • Voice

    Unknown

    No available official source states whether the free chat app offers a voice feature.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • chat.deepseek.com, plain fetch — HTTP 202 with an empty (0-byte) body, no text to read
    • chat.deepseek.com, browser fetch — returns DeepSeek's bot "Human Verification" challenge page ("Before proceeding to your request, you need to solve a puzzle…"), not the consumer chat product (captures/2026-09-15/src-deepseek-03.html, kept as history but removed from sources.tsv per the controller's ruling)
    • chat.deepseek.com, Wayback Machine — resolved a snapshot (20260915065927) but the decoded text was 8 characters, too short to use
    • DeepSeek Terms of Use (cdn.deepseek.com/policies/en-US/deepseek-terms-of-use.html) — no mention of a voice or speech feature
    • DeepSeek API docs (api-docs.deepseek.com/quick_start/pricing/) — describes text-based API models; no voice feature listed
    • Web search "DeepSeek chat app default model FAQ help.deepseek.com" — returned no help.deepseek.com or other DeepSeek-owned page describing a voice feature
    • DeepSeek app launch notice (api-docs.deepseek.com/news/news250115/, captured 2026-09-22) — no voice feature listed
    • DeepSeek App Store and Google Play listings (read 2026-09-22) — no voice feature named
    • TechNode, 14 September 2026 (press, read 2026-09-22): reports a gray test for some app users, a speaker button that reads replies aloud and four voice profiles, and says it "has not been announced as a full release"; no DeepSeek page announces it
    F-DEEPSEEK-FREE-TOOL-VOICE

  • Context window

    Confirmed

    DeepSeek's V4 release notice (24 April 2026) says a 1M-token context is now the default across all official DeepSeek services.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-CONTEXT-WINDOW-2

  • Context window

    Confirmed

    This is from DeepSeek's API documentation, not a statement about the free chat app. The pricing page states a context length of 1M (one million) tokens, shared by both API models.

    Interface:
    API
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-CONTEXT-WINDOW

  • Minimum age

    Not disclosed

    DeepSeek's Terms of Use do not publish a single minimum-age figure; they state the Services are "primarily intended for adults", and that users under 18 or under the minimum age required in their country must use the Services only with a legal guardian's consent.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-ACCOUNT-MIN-AGE

  • Phone requirement

    Confirmed

    DeepSeek's terms say an account is registered with an email or a third-party account, and its app notice lists the login methods as e-mail, Google Account and Apple ID. Neither names a phone number.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-DEEPSEEK-ALL-ACCOUNT-PHONE-2

  • Country availability

    Not disclosed

    DeepSeek publishes no supported-country list; the Terms say availability may vary by jurisdiction and bar use in comprehensively sanctioned countries without naming any.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Not established by this finding
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Not established by this finding
    • Egypt: Not established by this finding
    • Iraq: Not established by this finding
    • Jordan: Not established by this finding
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Not established by this finding
    • Morocco: Not established by this finding
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Not established by this finding
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not established by this finding
    • Tunisia: Not established by this finding
    • Yemen: Not established by this finding
    F-DEEPSEEK-ALL-COUNTRY-AVAILABILITY
Copilot11 findings

  • Default model

    Not disclosed

    Microsoft's help tells users to press "Quick response" under the prompt to choose another mode, and names no model for Quick response. It names GPT-5 for Smart mode only.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-ALL-DEFAULT-MODEL

  • Available models

    Confirmed

    Microsoft's help article on conversation modes says Smart mode uses the GPT-5 model and is available to all Copilot users. It says Think Deeper runs on "the latest reasoning models from OpenAI" without naming them, and names no model for Quick response.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-ALL-AVAILABLE-MODELS

  • Usage limit

    Not disclosed

    Microsoft describes the free Copilot plan as a "limited free version" subject to available capacity and limits, and says a user who hits a limit can upgrade to a paid option or come back later; the company publishes no number or time window for this limit.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-FREE-USAGE-LIMIT

  • Image generation

    Confirmed

    Microsoft's help pages say free Microsoft Copilot creates images. One page lists image creation among what signing in with a personal account gives; another says signing in gives more image creation.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Unknown

    Microsoft's help article for Microsoft Copilot documents uploading images and supported files into a chat, up to 20 files per conversation. No official page we collected says whether the free version includes it: the list of what "Microsoft Copilot (free) allows you to" do does not name uploads, and the evidence is for the product, not the plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • Microsoft Copilot for Young People / age limits (support.microsoft.com/.../microsoft-copilot-age-limits-and-parental-controls) — mentions uploading content to Copilot, but the page banner states "The information in this article applies only to the older version of the Microsoft Copilot app" (updated version as of August 18, 2026) and points web/updated-app users to "Microsoft Copilot for individuals: your data, privacy, and responsible AI", which was not captured
    • What's the difference between Microsoft Copilot (free) and Copilot in Microsoft 365 (support.microsoft.com) — free-section feature list does not mention file upload; "file upload" appears only under Copilot Chat in Microsoft 365 business plans
    • Getting started with Microsoft Copilot (support.microsoft.com) — no file-upload statement for the free plan
    • Copilot Pricing Plans for Individuals (microsoft.com/en-us/microsoft-365-copilot/pricing/individuals) — FAQ 08 says "Upload those files to Copilot" in a Microsoft 365 memory-transfer answer, not as a free-plan feature
    • AI credits and limits for Microsoft 365 subscriptions (support.microsoft.com) — paid-plan feature table, no free-plan file-upload statement
    • Microsoft help 'File upload in Microsoft Copilot' (captured 2026-09-22): product-level; names no plan
    F-COPILOT-FREE-TOOL-FILE-UPLOAD-2

  • Voice

    Confirmed

    The free Microsoft Copilot page says signing in with a personal account gives access to voice interactions.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-FREE-TOOL-VOICE

  • Context window

    Unknown

    Microsoft publishes no context-window figure (in tokens) for the free Copilot plan on any of the official pages reviewed.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • Copilot Pricing Plans for Individuals (microsoft.com/en-us/microsoft-365-copilot/pricing/individuals) — no context-window or token figures for any plan
    • What's the difference between Microsoft Copilot (free) and Copilot in Microsoft 365 (support.microsoft.com) — mentions "longer conversations" with sign-in but gives no token or context-window figure
    • AI credits and limits for Microsoft 365 subscriptions (support.microsoft.com) — feature-limit table has no token or context-window figures, and covers paid plans only
    • Getting started with Microsoft Copilot (support.microsoft.com) — no context-window figure
    • Conversation modes in Microsoft Copilot (support.microsoft.com, captured 2026-09-22) — modes and models, no context-window figure
    • Transparency Note for Microsoft Copilot (support.microsoft.com/en-us/privacy/microsoft-copilot/transparency-note), read 2026-09-22 — no context-window or token figure
    • File upload in Microsoft Copilot (captured 2026-09-22) — file count and size limits only
    F-COPILOT-FREE-CONTEXT-WINDOW

  • Minimum age

    Confirmed

    Microsoft's Copilot Supplemental Terms (effective 18 August 2026), which apply to the Copilot app on web, desktop and mobile, say a user must be at least 13, and sometimes older depending on the country's laws.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-ALL-ACCOUNT-MIN-AGE-2

  • Phone requirement

    Confirmed

    Copilot is signed into with a Microsoft Account, Apple ID or Google Account. Microsoft says a Microsoft account can be created with an email address instead of a phone number.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-COPILOT-ALL-ACCOUNT-PHONE-2

  • Country availability

    Confirmed

    Microsoft's official page says Copilot is available globally in over 170 markets, with exceptions for regions like China (excluding Hong Kong) and embargoed markets, without naming any of the 22 Arab League countries as either available or excluded.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Not established by this finding
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Not established by this finding
    • Egypt: Not established by this finding
    • Iraq: Not established by this finding
    • Jordan: Not established by this finding
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Not established by this finding
    • Morocco: Not established by this finding
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Not established by this finding
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not established by this finding
    • Tunisia: Not established by this finding
    • Yemen: Not established by this finding
    F-COPILOT-ALL-COUNTRY-AVAILABILITY
Claude12 findings

  • Default model

    Unknown

    Neither Claude's official pricing page, its Help Center, nor its release notes state which model is the default for Claude Free users; these sources list which models are available (Sonnet and Haiku) without saying which one is used by default.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • claude.com/pricing — plan feature and model comparison table (lists Free's available models, not a default)
    • Claude Help Center — 'How do usage and length limits work?' article
    • Claude Help Center — 'Release notes' page (September 2026 entries)
    • claude.com/pricing (captured 2026-09-22 after its redesign) — Free column lists Sonnet and Haiku as Yes, no default
    • Claude Help Center, 'Change the model, effort, and thinking settings' (8664678), read 2026-09-22 — explains switching models, names no default for Free
    F-CLAUDE-FREE-DEFAULT-MODEL

  • Available models

    Confirmed

    The plan-comparison table on Claude's pricing page shows Free-plan users can use the Sonnet and Haiku models, while Opus is not available to them.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-FREE-AVAILABLE-MODELS

  • Usage limit

    Confirmed

    The FAQ on Claude's pricing page states every plan, including Free, has usage limits that reset on a rolling five-hour session window, and that Free "covers everyday questions," with no fixed message count given.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-FREE-USAGE-LIMIT

  • Image generation

    Confirmed

    Claude's help center says Claude does not generate photos or illustrations the way image-generation tools do. It can draw diagrams and interactive visuals with HTML and SVG, and read uploaded images.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-ALL-TOOL-IMAGE-GENERATION

  • File upload

    Unknown

    Claude's help center documents uploading documents and images to chats, up to 20 files per chat. No official page we collected says whether the Free plan includes it: the pricing table has no upload row, and the evidence is for the product, not the plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • claude.com/pricing — Free plan feature list and plan-comparison table (only 'Create and edit files with code execution', not upload, is listed)
    • Claude Help Center — 'How do usage and length limits work?' article (no mention)
    • Claude Help Center — 'Upload files to Claude' article title seen only as a navigation link, its content not captured
    • Claude help 'Upload files to Claude' (support.claude.com/en/articles/8241126, captured 2026-09-22): product-level; names no plan
    F-CLAUDE-FREE-TOOL-FILE-UPLOAD-2

  • Voice

    Confirmed

    The plan-comparison table on Claude's pricing page shows Voice mode as Yes for the Free plan. Claude's help center says voice mode, a two-way spoken conversation, is a beta feature on all plans in the mobile apps, the desktop app and on the web.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-FREE-TOOL-VOICE

  • Context window

    Not disclosed

    Claude's pricing table gives no single context-window figure for the Free plan; it reads "Up to 1M, varies by model", the same wording as the paid plans.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-FREE-CONTEXT-WINDOW-2

  • Minimum age

    Confirmed

    Claude's Help Center states all users must be at least 18 years old to create and use a Claude account.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-FREE-ACCOUNT-MIN-AGE

  • Phone requirement

    Confirmed

    Claude's Help Center states phone verification is required for all new users when creating a Claude account, and there is no way to skip this step.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-CLAUDE-FREE-ACCOUNT-PHONE

  • Country availability

    Confirmed

    Anthropic's official Supported Countries and Regions page names 20 of the 22 Arab League countries among the countries, regions and territories where Claude.ai access is currently offered.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Supported
    • Bahrain: Supported
    • Djibouti: Supported
    • Algeria: Supported
    • Egypt: Supported
    • Iraq: Supported
    • Jordan: Supported
    • Comoros: Supported
    • Kuwait: Supported
    • Lebanon: Supported
    • Libya: Supported
    • Morocco: Supported
    • Mauritania: Supported
    • Oman: Supported
    • Palestine: Supported
    • Qatar: Supported
    • Saudi Arabia: Supported
    • Sudan: Supported
    • Somalia: Supported
    • Syria: Not established by this finding
    • Tunisia: Supported
    • Yemen: Not established by this finding
    F-CLAUDE-ALL-COUNTRY-AVAILABILITY-1

  • Country availability

    Confirmed

    Neither Syria nor Yemen appears on Anthropic's official list of countries, regions and territories where Claude.ai access is currently offered, a list presented as the complete set of currently supported countries.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    Country detail for this finding
    • United Arab Emirates: Not established by this finding
    • Bahrain: Not established by this finding
    • Djibouti: Not established by this finding
    • Algeria: Not established by this finding
    • Egypt: Not established by this finding
    • Iraq: Not established by this finding
    • Jordan: Not established by this finding
    • Comoros: Not established by this finding
    • Kuwait: Not established by this finding
    • Lebanon: Not established by this finding
    • Libya: Not established by this finding
    • Morocco: Not established by this finding
    • Mauritania: Not established by this finding
    • Oman: Not established by this finding
    • Palestine: Not established by this finding
    • Qatar: Not established by this finding
    • Saudi Arabia: Not established by this finding
    • Sudan: Not established by this finding
    • Somalia: Not established by this finding
    • Syria: Not supported
    • Tunisia: Not established by this finding
    • Yemen: Not supported
    F-CLAUDE-ALL-COUNTRY-AVAILABILITY-2
Grok11 findings

  • Default model

    Not disclosed

    Grok's plan table shows one model, Grok 4.6, in its Models row with a mark in the Free column, but no official page names a default model for the Free plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GROK-FREE-DEFAULT-MODEL

  • Available models

    Confirmed

    Grok's plan table marks the Free column in the Grok 4.6 model row. The table does tell yes from no: the Video generation row carries a faded mark in the same column.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GROK-FREE-AVAILABLE-MODELS

  • Usage limit

    Not disclosed

    xAI says the Free plan gives Grok "within generous limits" and that the free tier has its own limits on Chat and Voice that reset on their own schedule; it publishes no number or time window for them.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GROK-FREE-USAGE-LIMIT

  • Image generation

    Confirmed

    Grok's plan table marks the Free column in the "Image generation (Imagine)" row, while the "Video generation" row carries a faded mark in the same column: image generation is on the Free plan, video generation is not.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GROK-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Unknown

    Grok's apps FAQ documents file upload in chats on the web, iOS and Android, and says limits vary by platform or subscription. No official page we collected says whether the Free plan includes it: the evidence is for the product, not the plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • Grok apps FAQ (docs.x.ai/grok/faq, captured 2026-09-22): documents upload for the product; names no plan
    • Grok plan comparison table (x.ai/pricing, captured 2026-09-22): has no file-upload row
    F-GROK-FREE-TOOL-FILE-UPLOAD

  • Voice

    Confirmed

    Grok's apps FAQ says the free tier has its own limits on Chat and Voice, and the pricing page lists Voice mode under the Free plan: voice is on the Free plan, with an unpublished limit.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-GROK-FREE-TOOL-VOICE

  • Context window

    Unknown

    No available official source states the context window of the consumer Grok apps on the Free plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • x.ai/pricing (captured 2026-09-22): plan table lists models and features, no context-window figure
    • docs.x.ai/grok/faq and docs.x.ai/grok/overview (captured 2026-09-22): no context-window figure for the website or apps
    • API model pages (docs.x.ai/developers) give API context windows; not used, because the API is a different product from the consumer apps
    F-GROK-FREE-CONTEXT-WINDOW

  • Minimum age

    Confirmed

    SpaceXAI's consumer terms require users to be at least 13 years old or the minimum age required in their country, and users aged 13 to 17 need a parent's or guardian's permission.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-GROK-ALL-ACCOUNT-MIN-AGE

  • Phone requirement

    Unknown

    No available official source says whether a phone number is needed to create a Grok account.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • docs.x.ai/grok/faq (captured 2026-09-22): sign-in methods named are Apple, X, Google and email; no statement that a phone number is or is not required
    • x.ai/legal/terms-of-service (captured 2026-09-22): no phone-number requirement stated
    F-GROK-ALL-ACCOUNT-PHONE

  • Country availability

    Unknown

    xAI publishes no official list of the countries where Grok is available, and names no Arab League country.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • x.ai/pricing, docs.x.ai/grok/faq, docs.x.ai/grok/overview (captured 2026-09-22): no country list
    • x.ai/legal/terms-of-service (captured 2026-09-22): export-control and sanctions clause only, no country list
    • Web search for an official Grok country list (2026-09-22): only third-party pages and a February 2025 post from the @grok account about Google Play regions; not used as current evidence
    F-GROK-ALL-COUNTRY-AVAILABILITY
Perplexity13 findings
Mistral Vibe11 findings

  • Default model

    Unknown

    No available official source names the default model of Vibe's Free plan (Vibe is the new name of Le Chat).

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • mistral.ai/pricing (captured 2026-09-22): Free plan gives "our latest models", unnamed; the FAQ "Which model should I use?" names Mistral Medium for API tasks, not the app default
    • mistral.ai/products/vibe (captured 2026-09-22): no default model named
    • help.mistral.ai Vibe collection (2026-09-22): no article on the app's default model
    F-MISTRAL-VIBE-FREE-DEFAULT-MODEL

  • Available models

    Not disclosed

    Mistral says the Free plan gives limited access to Vibe "including our latest models", without naming them.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-MISTRAL-VIBE-FREE-AVAILABLE-MODELS

  • Usage limit

    Not disclosed

    Mistral says Vibe's Free plan gives "limited" messages and web searches, and that Pro gives "up to 6x" the free messages; it publishes no number for the Free plan. Mistral says accounts and plans moved unchanged when Le Chat was renamed Vibe.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-MISTRAL-VIBE-FREE-USAGE-LIMIT

  • Image generation

    Confirmed

    The Free plan card on Mistral's pricing page lists image generation, and the FAQ says Pro gives up to 40x more image generations than Free; no number is published for Free.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-MISTRAL-VIBE-FREE-TOOL-IMAGE-GENERATION

  • File upload

    Confirmed

    Mistral's pricing comparison reads "Limited" in the Free column of the "Libraries and document upload" row, against "Up to 15GB" for Pro.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-MISTRAL-VIBE-FREE-TOOL-FILE-UPLOAD

  • Voice

    Confirmed

    Mistral's pricing comparison puts the included mark in the Free column of the "Voice mode" row, and the table does tell yes from no (the "White label" row carries a different mark in the same column). The row describes recording or uploading audio, transcription, questions and summaries, and the Vibe page says voice input is powered by Voxtral. No official page says Vibe answers by voice.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Official source
    Last checked:
    F-MISTRAL-VIBE-FREE-TOOL-VOICE

  • Context window

    Unknown

    No available official source states the context window of Vibe's Free plan.

    Interface:
    All interfaces
    Plan:
    Free
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • mistral.ai/pricing and mistral.ai/products/vibe (captured 2026-09-22): no context-window figure for the app
    • Model pages on docs.mistral.ai give API context lengths; not used, because the app's model is not named
    F-MISTRAL-VIBE-FREE-CONTEXT-WINDOW

  • Minimum age

    Confirmed

    Mistral's help article (dated 12 August 2026) says users must be at least 13 years old, and a minor needs a parent's or guardian's permission to create an account.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Official source
    Last checked:
    F-MISTRAL-VIBE-ALL-ACCOUNT-MIN-AGE

  • Phone requirement

    Unknown

    No available official source says whether a phone number is needed to create a Vibe account.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • help.mistral.ai, all collections (2026-09-22): account articles cover email, password and 2FA, with no phone-number requirement stated
    • legal.mistral.ai/terms: connection timed out on 2026-09-22 (plain and browser)
    F-MISTRAL-VIBE-ALL-ACCOUNT-PHONE

  • Country availability

    Unknown

    Mistral publishes no official list of the countries where Vibe is available, and names no Arab League country.

    Interface:
    All interfaces
    Plan:
    All plans
    How to read this lens
    Not established by the available evidence
    Last checked:
    What was checked · notes in the original language
    • mistral.ai/pricing (captured 2026-09-22): a country selector for prices, not an availability list
    • help.mistral.ai, all collections (2026-09-22): no availability article
    • legal.mistral.ai/terms: connection timed out on 2026-09-22 (plain and browser)
    F-MISTRAL-VIBE-ALL-COUNTRY-AVAILABILITY

Observed changes (1) · Record corrections & improvements (25)

  • Claude · Observed change ·

    Claude's pricing table no longer shows 200k for the Free plan; it now reads "Up to 1M, varies by model".

    Previous claim · closed

    The plan-comparison table on Claude's pricing page shows the Free plan's context window as 200k tokens.

    Replacement finding: Claude's pricing table gives no single context-window figure for the Free plan; it reads "Up to 1M, varies by model", the same wording as the paid plans.

Record corrections & improvements (25)25

  • Copilot · Correction ·

    Recorded from a page about the older Copilot app, as the only official source. The Copilot Supplemental Terms (effective 18 August 2026), captured on 22 September 2026, state it for the current app.

    Previous claim · closed

    Microsoft's official page requires Copilot users to be at least 13 years old, with a higher minimum in some countries depending on regional requirements. The page states it applies to "the older version" of the Copilot app and points web/updated-app users to a different article that was not captured here; it is nonetheless the only official source found that publishes a minimum-age figure.

    Replacement finding: Microsoft's Copilot Supplemental Terms (effective 18 August 2026), which apply to the Copilot app on web, desktop and mobile, say a user must be at least 13, and sometimes older depending on the country's laws.

  • ChatGPT · Correction ·

    Recorded as publishing no figures for tool limits, file uploads included. The File Uploads FAQ, captured on 22 September 2026, publishes one for file uploads: 3 per day.

    Previous claim · closed

    File uploads, image generation, voice and data analysis on ChatGPT Free carry separate usage limits from the main chat limit; OpenAI does not publish figures for them.

    Replacement finding: On ChatGPT Free, image generation, voice, data analysis and other tools have usage limits separate from the main chat limit, and OpenAI publishes no figures for them. File uploads are the exception, with a published limit of 3 a day.

  • Perplexity · Correction ·

    Recorded as Confirmed unavailable, from the plans page alone. The image-generation help article, captured on 22 September 2026, says the Free plan gets limited image generations, so the cell is now Disputed.

    Previous claim · closed

    Perplexity says the Free plan has no access to image generation.

    Replacement finding: Perplexity's own pages disagree. The plans page (last modified 5 September 2026) says the Free plan has no image generation, and its table marks Image Generation No for Free. The image-generation help article (last modified 3 September 2026) says: "Free Plan: Get access to limited image generations." The plans page is newer, but the conflict stands.

  • Gemini · Correction ·

    Recorded as Unknown. The feature table on Gemini's limits page and the Deep Research help articles, captured on 22 September 2026, answer it within Deep Research.

    Previous claim · closed

    None of the captured official Gemini pages states whether a user without a Google AI plan can use a web-search tool inside Gemini Apps; the pages describe Google Search's 'AI Mode' integration only as a benefit for Google AI Pro and Ultra subscribers.

    Replacement finding: Gemini's feature-availability table marks Deep Research with a checkmark in the 'Without a Google AI plan' column, and Google's Deep Research help articles (computer, Android and iPhone) say Gemini includes Google Search as a source by default. The evidence covers web search within Deep Research only; no captured page names a web-search tool for ordinary chat. Google also says Deep Research may be unavailable to users without a plan during periods of high demand.

  • Perplexity · Correction ·

    Recorded as publishing no number or time window. The comparison table on the same plans page, captured on 22 September 2026, gives two for the Free plan: 3 Pro Searches a day and 1 Research query a month.

    Previous claim · closed

    Perplexity says the Free plan gives "practically unlimited" basic searches and a "very limited" number of Pro Searches; it publishes no number or time window.

    Replacement finding: Perplexity says the Free plan gives practically unlimited basic searches; it publishes no number for them.

  • Perplexity · Correction ·

    Recorded as available on every plan, from a page that documents the feature for the product and names no plan. The lens asks about the free plan, and this evidence does not answer it, so the cell is now Unknown for the free plan.

    Previous claim · closed

    Perplexity's help article explains how to start a voice conversation by tapping the voice icon in the iOS app; it does not tie the feature to a paid or free plan.

    Replacement finding: Perplexity's help article explains how to start a voice conversation in the iOS app. No official page we collected says whether the Free plan includes it: the plans page and the pricing hub do not mention voice, and the evidence is for the product, not the plan.

  • Grok · Correction ·

    Recorded as available on every plan, from a page that documents the feature for the product and names no plan. The lens asks about the free plan, and this evidence does not answer it, so the cell is now Unknown for the free plan.

    Previous claim · closed

    Grok's apps FAQ says Grok accepts many file types in chats on the web, iOS and Android, and that limits can vary slightly by platform or subscription; the page does not state this for the Free plan alone.

    Replacement finding: Grok's apps FAQ documents file upload in chats on the web, iOS and Android, and says limits vary by platform or subscription. No official page we collected says whether the Free plan includes it: the evidence is for the product, not the plan.

  • Claude · Correction ·

    Recorded as Unknown. The help article "Can Claude produce images?", captured on 22 September 2026, answers it.

    Previous claim · closed

    Neither Claude's pricing page nor its plan-comparison table nor the 'usage and length limits' article states an image-generation feature for the Free plan. The Help Center article titled 'Can Claude produce images?' was not captured with its full content; its title appeared only as a navigation link on other captured pages.

    Replacement finding: Claude's help center says Claude does not generate photos or illustrations the way image-generation tools do. It can draw diagrams and interactive visuals with HTML and SVG, and read uploaded images.

  • Copilot · Correction ·

    Recorded as Unknown. Microsoft's help article on creating a Microsoft account, captured on 22 September 2026, answers it.

    Previous claim · closed

    The official pages reviewed do not state whether a phone number is required to create an account for Copilot; the getting-started page describes signing in with a Microsoft Account, Apple ID or Google Account without mentioning a phone number.

    Replacement finding: Copilot is signed into with a Microsoft Account, Apple ID or Google Account. Microsoft says a Microsoft account can be created with an email address instead of a phone number.

  • Claude · Correction ·

    Recorded as Unknown. The help article "Upload files to Claude", captured on 22 September 2026, answers it.

    Previous claim · closed

    Claude's pricing page states the Free plan includes "Create files and execute code," which is a different capability from uploading files. None of the captured sources explicitly states whether file upload is available on the Free plan. The Help Center article 'Upload files to Claude' was not captured with its full content; its title appeared only as a navigation link.

    Replacement finding: Claude's help center describes uploading documents and images to chats, up to 20 files per chat. The article ties uploading to no paid plan.

  • Claude · Correction ·

    Recorded as available on every plan, from a page that documents the feature for the product and names no plan. The lens asks about the free plan, and this evidence does not answer it, so the cell is now Unknown for the free plan.

    Previous claim · closed

    Claude's help center describes uploading documents and images to chats, up to 20 files per chat. The article ties uploading to no paid plan.

    Replacement finding: Claude's help center documents uploading documents and images to chats, up to 20 files per chat. No official page we collected says whether the Free plan includes it: the pricing table has no upload row, and the evidence is for the product, not the plan.

  • Copilot · Correction ·

    Recorded as Unknown. Microsoft's help article on file upload in Microsoft Copilot, captured on 22 September 2026, answers it.

    Previous claim · closed

    The captured official pages do not confirm whether the current free Microsoft Copilot plan lets users upload files. The age-limits article mentions uploading content to Copilot, but it applies only to the older version of the app; the current-app article it points to was not captured.

    Replacement finding: Microsoft's help article for Microsoft Copilot (not Copilot for work) says users can upload images and supported file types into a chat, up to 20 files per conversation. The article ties the feature to no paid plan.

  • Copilot · Correction ·

    Recorded as available on every plan, from a page that documents the feature for the product and names no plan. The lens asks about the free plan, and this evidence does not answer it, so the cell is now Unknown for the free plan.

    Previous claim · closed

    Microsoft's help article for Microsoft Copilot (not Copilot for work) says users can upload images and supported file types into a chat, up to 20 files per conversation. The article ties the feature to no paid plan.

    Replacement finding: Microsoft's help article for Microsoft Copilot documents uploading images and supported files into a chat, up to 20 files per conversation. No official page we collected says whether the free version includes it: the list of what "Microsoft Copilot (free) allows you to" do does not name uploads, and the evidence is for the product, not the plan.

  • Copilot · Correction ·

    Recorded as Unknown. Microsoft's conversation-modes article, captured on 22 September 2026, names no model for the mode shown under the prompt.

    Previous claim · closed

    Microsoft does not name the underlying model that powers the free Microsoft Copilot plan on any of the official pages reviewed; the chat and image capabilities are described without naming a model.

    Replacement finding: Microsoft's help tells users to press "Quick response" under the prompt to choose another mode, and names no model for Quick response. It names GPT-5 for Smart mode only.

  • Copilot · Correction ·

    Recorded as Unknown. Microsoft's help article on Copilot's conversation modes, captured on 22 September 2026, names a model.

    Previous claim · closed

    Microsoft does not publish a list of named models available to free Copilot users; the official pages describe usage tiers ("Higher than free", "Extensive", "Highest") for paid plans without naming any specific model for any plan.

    Replacement finding: Microsoft's help article on conversation modes says Smart mode uses the GPT-5 model and is available to all Copilot users. It says Think Deeper runs on "the latest reasoning models from OpenAI" without naming them, and names no model for Quick response.

  • DeepSeek · Correction ·

    Recorded as Unknown. DeepSeek's app launch notice (January 2025), captured on 22 September 2026, lists web search among the free app's features.

    Previous claim · closed

    The captured pages do not state a web-search tool on the free chat app.

    Replacement finding: DeepSeek's app launch notice (15 January 2025) lists web search among the features of the free app on the App Store and Google Play. We found no newer official page that restates it.

  • DeepSeek · Correction ·

    Recorded as Unknown. DeepSeek's app launch notice and its App Store listing, captured on 22 September 2026, mention file upload.

    Previous claim · closed

    No available official source states whether the free chat app allows file uploads.

    Replacement finding: DeepSeek's App Store listing describes the app as free, and one of its version notes (July 2026) reads "Improved photo and file uploads". The app launch notice (January 2025) lists file upload and text extraction among its features.

  • DeepSeek · Correction ·

    Recorded as Unknown. DeepSeek's app launch notice, captured on 22 September 2026, lists the login methods, alongside the terms of use.

    Previous claim · closed

    DeepSeek's Terms of Use mention registering an account using "your Email or third-party account", without explicitly naming or ruling out a phone number as a general sign-up requirement, and the chat app itself could not be reviewed (see tried).

    Replacement finding: DeepSeek's terms say an account is registered with an email or a third-party account, and its app notice lists the login methods as e-mail, Google Account and Apple ID. Neither names a phone number.

  • Meta AI · Correction ·

    Recorded as Unknown. Meta's pages on the Meta Account and on creating a Facebook account, captured on 22 September 2026, answer it for meta.ai.

    Previous claim · closed

    The official Meta pages reviewed do not state whether a phone number is required to create the Meta account used with Meta AI; the available article only describes how a phone number or email is used after an account exists (login, notifications, verification), not a sign-up requirement.

    Replacement finding: Meta says Meta AI needs a Meta Account, and that Meta Accounts work on Facebook and Meta AI. Facebook's sign-up help asks for an email address or a mobile number, and confirms either one. This does not cover Meta AI inside WhatsApp.

  • Meta AI · Correction ·

    Recorded as Unknown. A WhatsApp help article captured on 22 September 2026 shows the list is not published, and two February 2025 press reports name Arab countries.

    Previous claim · closed

    Meta publishes no official page listing which countries Meta AI is or is not available in.

    Replacement finding: WhatsApp's help center says Meta AI is currently available only in "limited countries" and is rolling out gradually, without listing the countries.

  • Gemini · Correction ·

    Recorded as Unknown. Google's help article on creating a Google Account, captured on 22 September 2026, answers it.

    Previous claim · closed

    None of the captured official Gemini pages states whether a phone number is required to create or sign in to a Google Account used with Gemini.

    Replacement finding: Gemini needs a Google Account. Google's help on creating one lists adding a phone number as an optional step, and allows signing up with an existing non-Google email address.

  • Gemini · Correction ·

    Recorded as Unknown. Gemini's two app-store listings and the Gemini Live help articles, captured on 22 September 2026, answer it per platform.

    Previous claim · closed

    None of the captured Gemini help pages states whether voice chat (such as Gemini Live) is available to users without a Google AI plan; 'Gemini Live' appears only as an unlabelled row name in the Google AI plans comparison table, with no free-tier value extractable from the table's icon-based marks.

    Replacement finding: Gemini's Google Play listing names Gemini Live among the app's features without tying it to a paid plan, while it ties video creation to Google AI Plus, Pro and Ultra. Google's help article lists what Gemini Live needs (an Android device, the app, a Google account, age 13 or over) with no paid plan among them.

  • ChatGPT · Correction ·

    Recorded as Unknown. An OpenAI help article captured on 22 September 2026 answers it.

    Previous claim · closed

    The official OpenAI pages reviewed do not state whether a phone number is required to create a ChatGPT account generally; the available article only describes a phone-only sign-up path in specific countries, not the standard sign-up requirement.

    Replacement finding: OpenAI's help center says phone verification is no longer required to create a new OpenAI account or to use ChatGPT. It is still required when generating a first API key.

  • ChatGPT · Correction ·

    Recorded GPT-5.6 Terra among the Free models on every surface. The same comparison table limits it to Work and Codex on desktop, so it now has its own desktop record.

    Previous claim · closed

    ChatGPT's Free plan is marked Yes in the plan-comparison table for GPT-5.6 Luna and GPT-5 Thinking Mini, and gives GPT-5.6 Terra limited access restricted to ChatGPT Work and Codex on desktop.

    Replacement finding: ChatGPT's plan-comparison table marks GPT-5.6 Luna and GPT-5 Thinking Mini Yes for the Free plan.

  • Gemini · Correction ·

    Closed as a correction: the quoted rule is for Gemini Apps with a paid Google AI plan in the EEA, Switzerland or the UK (18+), not a free-plan minimum. Paid plans are out of vocabulary; the free/all minimum remains 13.

    Previous claim · closed

    Google states that using Gemini Apps with a Google AI subscription plan in the European Economic Area, Switzerland or the United Kingdom requires being 18 or older, instead of the usual 13-or-country-minimum that otherwise applies.