{"profile":"wrs-ai-tool-evidence-2","edition":"2026-10-28.1","checked_on":"2026-10-28","scope":"Public first-hand accounts and third-party tests, selected for useful methods and concrete trade-offs. WRS checked the linked reports, not the authors' accounts or billing records. They are not a representative survey, verified-purchase reviews or a WRS comparative benchmark.","cases":[{"id":"terra-visual-context-repairs","title":"Terra fixed small patches, but missed a character and a conversation detail","products":["chatgpt","claude"],"kind":"Named author’s task tests · paid training offered on the site","author":"Dustin’s AI Lab","date_label":"26 October 2026","surface":"GPT-6.1 Terra/Sol and Claude comparisons in a custom tool harness; not an ordinary Chat trial","method":"Synthetic review and screen tasks, with small repeated runs; three short bug patches were each replayed twice. A test with an accessible reference fix was discarded and rerun in isolation.","finding":"The author reports successful small repairs, but Terra missed a simplified/traditional Chinese character and misread a continuing conversation. Higher effort did not repair those screen-reading failures. Sonnet and Opus handled those two cases.","limits":"Small author-designed sample. The author also sells training; commercial independence is not established. Built-in Codex computer use and large patches were not tested. Do not infer a universal model order or consumer free-tier performance.","cost":"The report includes timings; a complete consumer-plan bill, allowance comparison and human repair cost are not established.","url":"https://www.agentcrew.cc/blog/posts/en/gpt-6-1-terra-review/","artifacts":[{"title":"Original Chinese version of the same report","url":"https://www.agentcrew.cc/blog/posts/zh/gpt-6-1-terra-review/"}]},{"id":"terra-api-effort-small-tasks","title":"More thinking cost more; it did not improve these short answers","products":["chatgpt"],"kind":"Commercial wrapper’s own API task test","author":"MultiAI","date_label":"Launch-period test, published 26 October 2026; checked 28 October","surface":"GPT-6.1 Terra through OpenAI’s API, not a ChatGPT subscription","method":"Eight short tasks with scripted right answers, each run three times at Off, Medium and Max. Examples include meeting actions, a multi-currency invoice and a date function.","finding":"The publisher reports 19/24 correct at Off and 24/24 at both Medium and Max. Its short-task sample supports trying a checked lower-effort setup before paying for more thinking; it does not establish equivalence on long projects.","limits":"MultiAI sells access to models. These are its measurements, not an independent WRS benchmark. The article describes the checks, but this guide does not certify the inputs, evaluator or account. App allowances and free chat were not tested.","cost":"Projected list-rate cost per 1,000 answers: US$1.63 at Medium and US$7.21 at Max, using its measured tokens. These are projections, not a purchased 1,000-answer bill, subscription charge or complete labour cost.","url":"https://multiai.co/blog/gpt-6-1-terra","artifacts":[]},{"id":"qianwen-doubao-shopping-oct23","title":"Shopping moved into the chat window; check the final basket","products":["qwen","doubao"],"kind":"Original Chinese newsroom hands-on inspection","author":"Li Yutong and Chen Ting, National Business Daily","date_label":"Published 23 October 2026; shopping observations dated 22 October","surface":"Mainland Qianwen and Doubao consumer apps; model and account plans unspecified","method":"Reporters inspected shopping flows, live-shop links and the same shop’s kitchen-paper offer through two channels.","finding":"They observed in-app selection and payment in Qianwen, and live-shopping links in Doubao. A paper offer cost CNY17.43 through Doubao after coupons versus CNY27.90 through the other channel on that date.","limits":"Selected accounts, offers and dates; no representative price or model-quality comparison. Coupon eligibility and checkout can change. The report’s company statements and expert forecasts are separate from these observations.","cost":"The quoted amounts are dated goods prices, not AI subscription fees. Current prices and savings for your account are unknown.","url":"https://www.nbd.com.cn/articles/2026-10-23/4614334.html","artifacts":[]},{"id":"everyday-five-prompts","title":"A party plan, a rewrite and a realistic evening routine","products":["chatgpt","claude"],"kind":"Editorial hands-on comparison · historical version","author":"Amanda Caswell, Tom’s Guide","date_label":"24 September 2026","surface":"ChatGPT-6 and Claude Opus 5.5; account tiers and settings not stated.","method":"Five shared prompts, with screenshots: party planning, rewriting, a holiday choice, public spending and household organisation.","finding":"The reviewer preferred Claude in four tasks, but ChatGPT for a manageable evening routine. This is a useful reminder to judge whether you can use the answer, not just its length.","limits":"One reviewer, selected prompts, no repeated trials or blinded scoring. Affiliate-funded publisher. This ChatGPT-6 result is not a test of GPT-6.1 Sol or of every current ChatGPT mode. We do not reuse the article’s unverified travel-cost arithmetic.","cost":"No task-level bill or free-quota measurement; a paid model result does not establish free-tier access.","url":"https://www.tomsguide.com/ai/i-tested-chatgpt-6-vs-claude-opus-5-5-with-5-everyday-prompts-it-wasnt-even-close","artifacts":[]},{"id":"claude-email-extra-usage","title":"Three email edits after the subscription limit","products":["claude"],"kind":"Public user self-report · billing not verified","author":"Reddit user Epsilonze","date_label":"22 October 2026","surface":"Claude Pro user reporting Opus 5.5 on High, after reaching a weekly limit.","method":"The user describes buying £5 of extra usage, then requesting and revising a short job-application follow-up email.","finding":"They report the credit disappearing during three short email requests. The post illustrates the surprise of moving from included usage to paid usage.","limits":"No invoice, full transcript, settings export or token trace. WRS cannot verify the charge or what else was processed. It is not a £5-per-email tariff; do not extrapolate nine emails to £15. Comments about caching and plan equivalence are not verified here.","cost":"£5 is the user’s claimed top-up, not a measured price for this task. Review extra-usage settings before continuing past a plan limit.","url":"https://www.reddit.com/r/claude/comments/1xcmf2r/3_messages_asking_for_an_email_for_a_follow_up_to/","artifacts":[]},{"id":"gemini-argon-daily-quota","title":"The same new Gemini, very different allowance stories","products":["gemini"],"kind":"Contrasting public user self-reports","author":"Reddit users Nug__Nug, Dingleberry_Blumpkin and Beautiful_Passion_17","date_label":"October 2026 discussion; exact individual posting dates not established","surface":"Gemini app, Gemini 4 Argon. Two commenters identify Pro; the study-help commenter does not establish a plan.","method":"Participants describe ordinary questions, a Google Sheets/skills task, and asking for explanations of thermodynamics course material.","finding":"One Pro user reports 52% of a five-hour allowance used by seven queries. Another reports 12% after three skills-and-ledger builds. A student describes useful explanations with much lighter apparent usage.","limits":"No reproducible prompts, account telemetry or verified purchases. These conflicting accounts cannot establish average capacity, a free allowance, or a diagnosed billing bug. Study help is not evidence that the generated explanations were correct.","cost":"Reported percentages are account observations, not money or transferable tasks per plan. Check your own allowance before and after your representative task.","url":"https://www.reddit.com/r/GeminiAI/comments/1x54x39/argon_4_thoughts/","artifacts":[]},{"id":"kimi-k31-usable-output","title":"Kimi K3.1 Pro: a working visualisation can still need repairs","products":["kimi"],"kind":"Named reviewer’s hands-on tasks","author":"象先志, published by TMTPost","date_label":"25 October 2026","surface":"Kimi K3.1 Pro, normal Pro speed rather than Highspeed; reviewer account tier not established.","method":"Seven coding tasks in fresh conversations, allowing self-checks before first delivery; examples include animation, a collision demo and an audio visualiser.","finding":"The reviewer reports that changing the audio input changed the displayed spectrum. A bicycle-and-bird scene missed the requested physical arrangement. They found normal Pro subjectively faster than K3.","limits":"Not a controlled speed measurement or an independent physics audit. Tests and output illustrations are the reviewer’s, not WRS runs. The article’s reported official benchmarks and other users’ quota claims are not used as measured results here.","cost":"No bill for the reviewer’s seven tasks. API output tariffs and other people’s plan percentages cannot price an app project.","url":"https://www.tmtpost.com/8180725.html","artifacts":[]},{"id":"kimi-k31-effort-repairs","title":"A good-looking Kimi build is not the end of the task","products":["kimi"],"kind":"Named user’s app and coding-tool experience","author":"まさお @未経験からプロまでAI活用","date_label":"24 October 2026","surface":"Kimi K3.1 Pro: Kimi Code on the author’s Max plan, plus separate app experiments.","method":"A 25-question custom benchmark; only 15 questions were scored within the available allowance. The author also tried a character-to-3D app task.","finding":"They liked landing pages and an editor’s undo/redo, but found a marble run failed under repeated use. Maximum effort interrupted benchmark work; they preferred starting at high effort.","limits":"Personal benchmark and judgement, not representative usage. The Code allowance observation is not a free-chat quota. Account naming may reflect a legacy plan. App and Code results are kept separate; no claim that every app offers the same effort control.","cost":"The author reports a Max subscription, without an itemised task bill. Their stopped benchmark does not establish how many free projects another account can finish.","url":"https://note.com/masa_wunder/n/n47097866e600","artifacts":[]},{"id":"sol-web-brief-version","title":"Sol, Opus and Grok tried the same small website brief","products":["chatgpt","claude","grok"],"kind":"Named reviewer’s task comparison · tool surface","author":"Thomas Wiegold","date_label":"5 October 2026","surface":"GPT-6.1 Sol, Claude Opus 5.5 and Grok 4.7; this is a coding workflow, not a free consumer-chat test.","method":"One shared design brief and two Go algorithm tasks. The article links the design outputs for inspection.","finding":"The reviewer preferred Opus’s website, called Sol’s clean but underdeveloped, and placed Grok between them. This supports inspecting the delivered page against the brief rather than treating a model label as a result.","limits":"One author, subjective design preference, limited tasks. No inference about ordinary chat quality or all projects. API-price comparisons in the article do not measure subscription costs.","cost":"The author says they pay for Claude Max and ChatGPT’s $20 plan; no itemised cost per tested build.","url":"https://thomas-wiegold.com/blog/gpt-6-1-sol-review/","artifacts":[]},{"id":"doubao-game-facts","title":"When a game guide keeps mixing up characters","products":["doubao"],"kind":"App Store user complaint · historical, version unknown","author":"China App Store reviewer 77和笨鸟","date_label":"4 September shown; year and app version not established","surface":"Doubao consumer app; model, plan and search settings not supplied.","method":"The reviewer says they asked about character attributes in the game 代号鸢 and repeatedly corrected confused character names.","finding":"They report that the corrections did not reliably repair later answers. A practical lesson is to provide the official character page and ask for a source-linked table rather than keep asking the same unsourced question.","limits":"No transcript, preserved prompts or independently checked outputs. The displayed date cannot be attached to the newest model. A selected complaint is neither a failure rate nor proof that the current app has the same fault.","cost":"No payment or allowance observation. Paying is not established as a fix for this reported factual problem.","url":"https://apps.apple.com/cn/app/%E8%B1%86%E5%8C%85-%E7%94%9F%E6%B4%BB%E5%B7%A5%E4%BD%9C-ai-%E5%8A%A9%E6%89%8B/id6459478672","artifacts":[]},{"id":"everyday-paid-use","title":"Living with three paid assistants","products":["chatgpt","claude","gemini"],"kind":"Named user's experience","author":"Learn With Darin","date_label":"Updated September 2026; individual session dates not supplied","surface":"Consumer subscriptions; several model versions, not a frozen comparison","method":"The author reports paying for all three and using them daily for over a year. No matched outputs or task counts accompany the preferences.","finding":"Darin prefers Claude for prose, ChatGPT for structured writing and breadth, and Gemini for editing inside Google Docs. Those are personal preferences, not measured win rates.","limits":"The author declares no affiliation with the three providers. The page predates October releases; its model and feature list is not our current product specification.","cost":"Paid consumer use is reported; no invoices or per-task costs are published.","url":"https://learn.techwithdarin.com/compare/chatgpt-vs-claude-vs-gemini/","artifacts":[]},{"id":"devops-runtime-check","title":"Good-looking code that would not start","products":["chatgpt","claude","gemini"],"kind":"Third-party task test","author":"Josphat Mutai / ComputingForGeeks","date_label":"Tested 13 October 2026","surface":"Gemini 4 Argon via OpenRouter; GPT-6.1 Sol and Opus 5.5 via their APIs","method":"Three infrastructure tasks, repeated three times. The author ran the generated server configuration, rather than only checking its syntax.","finding":"Gemini's initial server configurations failed to start; Sol and Opus passed that check. Gemini repaired all three retested failures after receiving the error, but two repairs changed a requested image or port.","limits":"One publisher, tiny task set, different routes and effort settings; no inference about the consumer apps or overall reliability. A code excerpt and runtime error are inspectable, not an independently replicated full benchmark.","cost":"Reported nine-call API charges: Sol $0.062, Opus $0.254, Argon default $0.045 or high $0.169. These exclude the person's time and infrastructure, and are not monthly app prices.","url":"https://computingforgeeks.com/gemini-4-argon-released-features-benchmarks/","artifacts":[]},{"id":"four-player-game","title":"An existing game gains four-player support","products":["claude"],"kind":"Named developer's working example","author":"Simon Willison","date_label":"24 October 2026","surface":"Claude Opus 5.5; subscription and execution settings not specified","method":"The developer publishes his starting prompt and links the resulting game release and replay.","finding":"Willison reports extending Battledrones from two to four players, including a new map and multiplayer changes. The linked release gives readers an output to inspect.","limits":"An experienced developer working in an existing project. No comparison group, full correction count or independent security assessment; WRS has not run the game. This report does not establish a beginner's expected success rate.","cost":"Task price, token usage and total human time are not disclosed.","url":"https://simonwillison.net/2026/Oct/24/battledrones/","artifacts":[{"title":"Published game release","url":"https://github.com/simonw/battledrones/releases/tag/v14"}]},{"id":"deepseek-chinese-checks","title":"Chinese short tasks passed; a money parser did not","products":["deepseek"],"kind":"Third-party small-sample test","author":"TOKRACE","date_label":"Text tests 18 October; visual follow-ups 20–21 October 2026","surface":"DeepSeek V4.1 Flash 1017, direct API; thinking on/off","method":"Fifteen Chinese text tasks and three JavaScript tasks in each mode. The publisher describes parameters and retains failures; supplementary visual request records are downloadable.","finding":"Both modes passed 15/15 text checks and 2/3 code checks. Money parsing failed through exhausted output budget or incorrect handling of leading zeros. Visual follow-ups also retained connection failures and retries.","limits":"Publisher-reported results, not a WRS replication. Small tasks, mutable model alias and differing visual retry budgets prevent a general ranking. Connection failures are not all model-quality failures. Commercial independence is not established.","cost":"Do not transfer the short-test cost to a full app. This guide has no verified invoice for these runs.","url":"https://www.tokrace.com/zh-CN/guides/deepseek-v4-1-flash-1017","artifacts":[{"title":"Supplementary request and failure records","url":"https://www.tokrace.com/evaluations/deepseek-v4-1-flash-1017-supplement.json"}]},{"id":"deepseek-cache-bill","title":"A low bill with an important catch: cached input","products":["deepseek"],"kind":"Public user self-report","author":"Reddit user illumihani","date_label":"Read 25 October 2026; exact posting date not established","surface":"V4.1 Flash 1017 used on a software project, not the free chat app","method":"A public post includes usage screenshots and describes a week's work. The account, invoice and output quality have not been independently verified.","finding":"The user reports about six billion tokens and $46 over a week, mostly cache hits. They liked bounded tasks and test iteration, but reported invented details and skipped instructions; a security review was still pending.","limits":"Cached input means previously supplied text was reused at a different rate. This is one selected anecdote, not a normal bill or proof that the software is safe. Other tools and human work may add costs.","cost":"$46 is the user's reported API spend, not a verified all-in project cost or a budget promise.","url":"https://www.reddit.com/r/DeepSeek/comments/1xe9mw9/21b_tokens_in_a_single_day_for_16_v41_flash_1017/","artifacts":[]},{"id":"claude-quota-interruption","title":"A paid plan can still interrupt a session","products":["claude"],"kind":"Public user self-report","author":"Reddit user Silly-Catch-4518, in a community discussion","date_label":"Thread labelled 9 October 2026; comment date not independently established","surface":"Claude Pro; exact model and usage history not supplied","method":"One comment in a volunteer-run problem-report thread describes being stopped during project planning.","finding":"The user reports hitting a usage limit despite paying for Pro. This is a concrete purchase concern to test against your own workload.","limits":"No account logs or reproduction. A complaints thread selects for problems and cannot tell us how common they are, what caused this case, or whether it persists today.","cost":"The current official Pro monthly list price is $20 before applicable tax; paying does not remove all usage limits.","url":"https://www.reddit.com/r/ClaudeAI/comments/1x1ghzo/performance_and_bugs_discussion_hub_updated_on_9/","artifacts":[]},{"id":"metr-time-vs-feeling","title":"Measure finished work, not the feeling of speed","products":["chatgpt","claude","gemini","deepseek"],"kind":"Independent research; historical context","author":"METR","date_label":"Study used February–June 2025 tools; follow-up 24 February 2026","surface":"Experienced open-source developers, mainly Cursor with Claude 3.5/3.7 in the original study","method":"The original randomized study involved 16 developers and 246 tasks. METR later explained why selection and time-measurement problems limited its newer experiment.","finding":"The original tasks took 19% longer with AI, despite participants feeling faster. The follow-up says newer tools likely help more, but its data cannot reliably establish the size of that change.","limits":"This is a lesson about measuring productivity, not evidence that today's four products slow you down. It does not rank them or describe beginners' tasks.","cost":"Subscription price alone does not measure time saved. Track checking and repair time as part of your own trial.","url":"https://metr.org/blog/2026-02-24-uplift-update/","artifacts":[{"title":"Original randomized study","url":"https://arxiv.org/abs/2507.09089"}]},{"id":"grok-everyday-five","title":"Grok: a natural message, an incomplete shopping list","products":["grok","chatgpt"],"kind":"Publisher hands-on comparison","author":"Amanda Caswell · Tom’s Guide","date_label":"24 October 2026","surface":"Consumer chats; exact model versions and paid plans not stated.","method":"Five shared everyday prompts, including a short family message and a constrained grocery plan; selected screenshots.","finding":"The reviewer preferred Grok’s message tone but found assumed ingredients and unrealistic quantities in its meal plan.","limits":"Small subjective sample; no repeat runs or independent scoring. The publisher discloses affiliate links. Its overall winner is not a general product verdict.","cost":"Task charges and subscription plans not disclosed.","url":"https://www.tomsguide.com/ai/i-tested-chatgpt-vs-grok-with-5-questions-boomers-actually-ask-and-the-winner-surprised-me","artifacts":[]},{"id":"qianwen-doubao-shopping","title":"Shopping help: useful links still need a human quantity check","products":["qwen","doubao"],"kind":"Original Chinese newsroom hands-on test · historical","author":"Chen Ting and Li Yutong · National Business Daily (每日经济新闻)","date_label":"18 June 2026","surface":"Mainland Qianwen and Doubao apps plus shopping integrations, not international Qwen Chat.","method":"Reporters tried a travel-outfit brief and a two-person household shopping list.","finding":"Qianwen supplied product links but still needed manual selection. Doubao’s price comparisons were useful; its household quantities were excessive.","limits":"Historical features, no repeated benchmark or disclosed model versions. Store coverage and promotions affect results; these observations do not establish current best prices.","cost":"App-plan charges and time per task were not disclosed; shopping totals are not subscription prices.","url":"https://www.nbd.com.cn/articles/2026-06-18/4430941.html","artifacts":[]},{"id":"kimi-k3-repair-quota","title":"Kimi K3: an attractive prototype needed repair and more allowance","products":["kimi"],"kind":"Publisher hands-on report · historical model","author":"Photon Planet (光子星球), credited original report reproduced by Sina","date_label":"18 July 2026","surface":"K3 work/coding tasks; report explicitly identifies Kimi Code for its pricing-analysis task. Not a current consumer-chat comparison.","method":"Seven custom tasks included an interactive 3D portfolio, requirements checks and debugging, with prompts and screenshots.","finding":"The portfolio first opened blank, then ran after repair; quota pressure prevented further mobile fixes.","limits":"One publisher’s task set, unknown repetitions and commercial relationship. Code/Work results do not establish K3.1 Pro app performance.","cost":"The portfolio reportedly took nearly an hour. Exact task bill and tokens are not supplied; historical plan claims are not current checkout prices.","url":"https://finance.sina.cn/stock/jdts/2026-07-18/detail-iniiefhq1428043.d.html","artifacts":[]},{"id":"doubao-app-review-limits","title":"A Doubao user reports losing instructions across turns","products":["doubao"],"kind":"Individual public app review · unverified self-report","author":"App Store reviewer ren ny (displayed as ren ny)","date_label":"App Store displays 5 July; year and exact app version are not shown.","surface":"Reviewer describes Doubao mobile and web use; plan not stated.","method":"A written account of drafting and question-answering; no preserved prompts, outputs or measured timing.","finding":"Reports peak-time delays, forgotten formatting constraints and repeated corrections.","limits":"Selected complaint, not a representative survey, verified purchase or measured failure rate. App updates may change this experience.","cost":"Payment status and task costs are unknown.","url":"https://apps.apple.com/cn/app/%E8%B1%86%E5%8C%85-%E7%94%9F%E6%B4%BB%E5%B7%A5%E4%BD%9C-ai-%E5%8A%A9%E6%89%8B/id6459478672","artifacts":[]},{"id":"kimi-k31-highspeed-quota-report","title":"One Kimi user ran into a Highspeed allowance","products":["kimi"],"kind":"Public user self-report","author":"xiazhi88 on LINUX DO","date_label":"23 October 2026","surface":"Kimi K3.1 Pro, including Highspeed; the writer mentions a CNY699 K3 plan but does not clearly establish the plan for the later K3.1 test.","method":"The writer describes continuous use and then trying Highspeed. Exact prompts, task outputs, total tokens and a receipt are not provided.","finding":"The user reports that a five-hour Highspeed allowance lasted less than two hours of their continuous use. They also describe earlier K3 weekly limits running out.","limits":"One self-report, not a controlled test, verified bill or representative account. After this post, Kimi said it adjusted settings; IT Home reproduced the 26 October announcement. The effect has not been independently checked here. This report does not establish free capacity or current behaviour.","cost":"The writer mentions a CNY699 plan for earlier K3 use; the K3.1 task bill and current plan are not established. Check the actual checkout and usage meter.","url":"https://linux.do/t/topic/3067215","artifacts":[{"title":"IT Home: Kimi statement and subsequent adjustment, 26 October","url":"https://www.ithome.com/1/017/723.htm"}]}],"recommendations":[{"product":"chatgpt","try_if":"You want one place to explore an idea, shape a draft and turn a small set of files into useful work. It is a sensible first trial for someone who has not yet settled on one workflow.","tradeoff":"Terra’s new reports cover a custom tool harness and the API. The screen test retained mistakes; the short API sample gained no accuracy from Max over Medium. These do not establish free Chat performance or long-project reliability. Inspect the actual model and check the output.","buy_when":"Start free for occasional drafting and explanations. Consider Plus at the listed US$20/month only when a feature or allowance you actually need interrupts recurring useful work. Track a few representative sessions, including corrections. Work/Codex shares an allowance; buying API credit does not expand an app plan. Do not pay just to obtain a model name.","check":"For everyday Chat, turn your rough email notes into a 120-word message and check names, dates and promises. For Work/Codex, if already available, use a fictional two-line receipt and ask for quantities, line totals and unknown charges. Compare every cell with the original before trying a higher effort setting. These are suggested checks, not WRS trials. Terra and Sol are not ordinary Chat models.","evidence":["terra-visual-context-repairs","terra-api-effort-small-tasks","sol-web-brief-version","everyday-five-prompts","everyday-paid-use","devops-runtime-check"]},{"product":"claude","try_if":"You spend time rewriting prose, assembling a document or iterating on a small interactive tool. Test whether its editing experience reduces the work between a first draft and something you can use.","tradeoff":"A small new screen-task report favoured Sonnet and Opus on two details that other tested models missed. Its author offers paid training; it is not a general model ranking. The email extra-usage story is one account. Longer answers and higher effort do not automatically improve a document; check facts and keep the final send under your control.","buy_when":"For occasional emails, begin with free access and a short input. Pro is US$20 billed monthly; the annual offer is $200 paid up front. Pay when a recurring document or writing workflow works for you and the included allowance is the actual bottleneck. Extra usage is a separate decision: set a spending limit rather than assume a top-up behaves like another month of Pro.","check":"Everyday task: draft a follow-up to a job application. Give the role, submission date, recipient and the one question you need answered; request a short, polite draft without invented qualifications. Edit it once for your own voice, then send it yourself. Before extra usage, inspect the limit and spending settings. A paid email anecdote is not a fixed cost per email.","evidence":["terra-visual-context-repairs","claude-email-extra-usage","everyday-five-prompts","everyday-paid-use","four-player-game","claude-quota-interruption","devops-runtime-check"]},{"product":"gemini","try_if":"Your work already lives in Google apps. The useful question is whether editing or finding information inside that workflow saves copying and switching between tabs.","tradeoff":"Current Argon user reports disagree sharply on allowance consumption. They are paid or unspecified accounts, not a measurement of free use. Google-app integration, stronger-model access and ordinary standalone chat have different conditions. Check the output and your own usage meter.","buy_when":"Stay free when short explanations and occasional drafts meet your need. The checked US listing is $4.99/month for AI Plus and $19.99 for AI Pro; confirm model and integration eligibility before buying. Pro/Ultra advertise Gemini 4 Argon access. Pay for a demonstrated missing feature or allowance, not the assumption that every paid session will be unlimited.","check":"Everyday task: ask it to explain one page of notes you are allowed to share. Request a plain-language explanation, one worked example and three quiz questions with separate answers. Check each answer against the notes. Compare your allowance before and after this session. Try a Google-app connection separately only if your account offers it; a standalone answer does not prove the connection works.","evidence":["gemini-argon-daily-quota","everyday-paid-use","devops-runtime-check"]},{"product":"deepseek","try_if":"You want a free chat starting point or can define small Chinese-language extraction and coding tasks with clear checks. Developers can separately explore the API, with a spending cap and inspectable outputs.","tradeoff":"The company reported the web/app update; V4.1 tests and the low-bill story in this guide use the API. They do not establish the performance, allowance or reliability of your own chat session. Check the model if your app displays it; otherwise treat it as unknown. Verify search claims against original sources.","buy_when":"The official assistant has free entry. We have not established a consumer subscription tariff or a fixed free-project allowance. Continue free when it completes your small tasks. Developer API access is a separate usage-billed route, useful for software you intentionally build; do not buy it merely because a chat answer failed or a server was busy.","check":"Everyday task: paste the public descriptions of three products you are comparing. Ask for a table of price, features, unknowns and a source for every claim; forbid filling gaps from memory. Open each cited page yourself. Try one follow-up that changes your budget. Keep any unsupported claim marked unknown rather than buying an API key to make the answer seem more authoritative.","evidence":["deepseek-chinese-checks","deepseek-cache-bill"]},{"product":"grok","try_if":"Try a concise everyday explanation or message. The recent external task comparison gives you concrete outputs to inspect.","tradeoff":"A conversational style you enjoy is useful, but is not a fact-checking guarantee. The latest everyday case has unknown model/settings; the explicit Grok 4.7 review is a coding comparison. Neither supplies a fixed free-message count. Verify linked claims, quantities and any additional spend.","buy_when":"Try the $0 entry first. The current US web listing starts at $8/month for Plus, with a shared weekly allowance across included Grok/Cursor activities; API usage is billed separately. Pay only when a tested recurring task needs the extra access. Check the offer name, renewal, remaining pool and top-up settings in your own account; older SuperGrok descriptions may differ.","check":"Everyday task: write a tactful message declining an invitation. Supply your real reason, desired tone and maximum length; request two options without making promises for you. Choose and edit one yourself. For a shopping list, give quantities, budget and exclusions explicitly and check every row. The recent everyday review found different strengths across these tasks; it did not establish the exact current model or free quota.","evidence":["sol-web-brief-version","grok-everyday-five"]},{"product":"qwen","try_if":"Try Chinese drafting or a small document task. Choose the mainland app when you specifically need its local integrations.","tradeoff":"The newer shopping inspection covers mainland app checkout; the June case remains historical. Compare the final basket, seller and coupon conditions. Neither establishes document accuracy, a fixed free PPT count or international Qwen performance. Check your own subscription page.","buy_when":"Start with free entry. For mainland Qianwen office work, the checked China listing includes CNY19, CNY49 and CNY128 monthly tiers; creative subscriptions and credits are separate. Pay after a useful outline/export workflow actually reaches an allowance or feature barrier. The membership agreement has rolling and monthly limits; spare monthly capacity does not override a shorter-window limit.","check":"Everyday task: in mainland Qianwen, turn your own five meeting bullets into a five-slide outline. Ask it to preserve the numbers, separate decisions from suggestions and leave missing owners unknown. Approve the outline before using a PPT tool; open the exported file and check text overflow and editable content. International Qwen has a separate interface and entitlements. This is a suggested check, not a new measured success. For a shopping suggestion, inspect the final basket without paying; check quantity, seller and coupon eligibility.","evidence":["qianwen-doubao-shopping-oct23","qianwen-doubao-shopping"]},{"product":"kimi","try_if":"Try a brief, slides or a small prototype with a clear output to inspect.","tradeoff":"K3.1 Pro has reports of useful builds and repairs. One user reported rapid Highspeed allowance use on 23 October. On 26 October Kimi said it had adjusted settings after usage complaints; we have not verified the effect. Plans share a monthly pool and Code has a rolling limit. App and Code evidence differ; one account cannot establish your free capacity.","buy_when":"Begin with the free allowance for one small draft and its repair. The current domestic monthly table starts at CNY49 Go; Code starts at CNY99 Plus. Pay when the correct product surface and a useful repeatable task need more access. K3.1 Pro reviewers still encountered repair and allowance limits. Include revision attempts in your budget; a higher tier is not a guaranteed finished website.","check":"Everyday task: create one local page for a club event from your own public date, location and three activities. Ask for a single HTML file, with no booking or publishing. Open it and change the date filter, then request one repair with the observed fault. Keep the working file before asking for visual polish. In Code, use a higher effort only for a specific remaining defect; do not assume the app has identical controls.","evidence":["kimi-k31-usable-output","kimi-k31-effort-repairs","kimi-k3-repair-quota","kimi-k31-highspeed-quota-report"]},{"product":"doubao","try_if":"Try an everyday Chinese question, a voice interaction or a small draft.","tradeoff":"Paid capacity does not verify facts. The new shopping inspection is a dated channel observation, not a promise of the lowest price. Check quantities and the final basket. Current official help describes model 2.1 Pro, shared advanced usage, and five-hour and seven-day limits; free capacity remains task-dependent.","buy_when":"Ordinary quick chat, search and writing have free entry. Try a small task first; advanced tools have daily trial limits. Standard is CNY68/month on recurring billing, not the separate CNY80 one-off month or CNY688 annual payment. Pay for recurring advanced work that is already useful. Standard does not buy extra video generation beyond its daily trial; check whether you actually need a higher tier.","check":"Everyday task: organise a game-night plan from your own public rules, budget and guest count. Ask for a short schedule and a shopping table, keeping missing prices unknown. Recalculate the total and check any game facts against their official page. If you need slides, approve the outline before using AI PPT and inspect the export. Save the draft before requesting another version.","evidence":["qianwen-doubao-shopping-oct23","doubao-game-facts","qianwen-doubao-shopping","doubao-app-review-limits"]}]}