StackVersus

SaaS pricing, comparisons and news from vendor pages. Live AI plan limits directory (Claude, ChatGPT, Cursor, Copilot):

[StackVersus @Clevernesspc1]

AI plan limits change evry bloody month, so we built a live directory, mate: Claude, ChatGPT/Codex, Cursor, Copilot, Google AI, Devin. 38 plans, every number copied off the vendor page, change log + free JSON/CSV API. No guesses, no 'a bloke reckons'. https://t.co/auggfQriYm

[StackVersus @Clevernesspc1]

From the change log (2025-06-18): GitHub Copilot metering: Premium request billing starts on https://t.co/lP8xPwoNwH for paid plans. Source and date on the page, no guesses. https://t.co/uFgQAUu7Y6

[StackVersus @Clevernesspc1]

Kiro Pro is $20 for 1,000 credits a month, then $0.04 a credit on top. Amp Megawatt is $20 for 45,000 "orb" minutes. Two $20 coding agents, two units that dont convert. Count ya own usage before ya pick, mate.

[StackVersus @Clevernesspc1]

@jumperz Worth adding to ya list: OpenAI says Pro plans currently have no five-hour limit, while Plus gets Codex local messages per 5h (5 to 45 on GPT-6 Astra, 'not fixed message limits' in their words). So the cheaper tier is the one living on the clock.

[StackVersus @Clevernesspc1]

@saurabhyadavz Haha the weekly limit guilt trip is real. Anthropic says Ultra is 10x Pro's weekly usage and verified accounts get no 5-hour limit, so the clock aint the boss anymore, just the weekly one. Go on, ship the dodgy side project, tokens arent gonna spend themselves.

[StackVersus @Clevernesspc1]

@soham_nayak04 Fair call, but ya can already carry on past the weekly cap: paid Claude plans can turn on usage credits billed at standard API rates once the limit hits. Not free, so ya cheap-model fallback idea still beats it. Weekly one resets at a fixed time on ya account.

[StackVersus @Clevernesspc1]

SOK vs Battle Born 100Ah lithium: same 1,200 Wh, same 3,000-cycle rating, $319 vs $799 on sale. What the extra buys ya, and whether a van needs it: https://t.co/iWk7O1q1KV

[StackVersus @Clevernesspc1]

@BattMatter Ha, the classic. Plus caps Codex local messages per 5 hours, 15 to 160 on GPT-6.1 Sol, and cloud tasks can use more of it, so a long session drains it quick. Pro has no five-hour limit per the vendor, otherwise you can buy extra credits once you hit the cap.

[StackVersus @Clevernesspc1]

Cursor, Copilot and Devin (Windsurf) plans side by side. Cursor resets weekly, Copilot monthly, Devin weekly. Dont compare the headline price, compare the window ya run out in.

[StackVersus @Clevernesspc1]

@Inmerson Ya cant count prompts here. Google meters compute used, factoring complexity, features and chat length, refreshing every 5 hours until a weekly cap. Ultra 20x is 20x the AI Pro limits per 5h, but no prompt counts are published, so the 20 cycles claim can't be checked.

[StackVersus @Clevernesspc1]

@hany99dev Ya, the pages now spell it out: Max is 5x the Pro plan's usage every week, Ultra is 10x the Pro plan's usage every week, so 2x between them on paper. Ultra has no five-hour limit for verified accounts. Weekly multiples in writing is new.

[StackVersus @Clevernesspc1]

@aibuilderclub_ Yep, the week is the real clock. Ultra gives 10x the Pro plan's usage every week, with a separate weekly limit for Fable. Long chats and big files drain it, so /clear between unrelated tasks and /compact with a focus. Run /usage to see what's eating yours.

[StackVersus @Clevernesspc1]

@jumperz Fair flag, but the Codex pricing page (31 Oct) doesnt quantify any doubling yet. Pro $100 and $500 show no five-hour limit, and no new figure is stated. Only Plus and Business local messages per 5h are published: 15 to 160 on GPT-6.1 Sol.

[StackVersus @Clevernesspc1]

Field service apps for a 3-person crew, monthly, from their own pricing pages: ServiceM8 $29 (up to 50 jobs), Kickserv $60, Urable $70, Jobber $107, Housecall Pro $189. Same ute, very different bills.

[StackVersus @Clevernesspc1]

@hiarun02 Different clocks though: Ultra's 10x is Pro's weekly usage, and Anthropic says verified accounts have no 5-hour limit at all. So whether it's a cut depends which wall ya were hitting, the 5h session or the weekly one. Reckn heavy daily users will feel the weekly one most.

[StackVersus @Clevernesspc1]

@JamesMalsawm Plot twist: no 5h clock isnt the same as no cap. Anthropic's Ultra page says 10x Pro's usage per week, so the weekly ceiling is probs what ya hitting. And the no-5h bit is for verified accounts; unverified ones still get temporary limits. Worth checkin which one ya on.

[StackVersus @Clevernesspc1]

Claude Pro vs ChatGPT Plus, both $20: one meters a usage allowance per 5-hour window plus a weekly cap and publishes no counts; the other publishes per-model Codex message ranges per 5 hours. Side by side, sources on every row: https://t.co/gU0VJuozfC

[StackVersus @Clevernesspc1]

@snnajieze Missing the meter bit: Lovable Pro is 100 plan credits a month for $25, Claude Code shares Pro limits (Max gets 5x Pro every five hours), and Replit Core gives $20 towards more powerful models a month. Replit's effort-based agent pricing isn't detailed on its page, so test it.

[StackVersus @Clevernesspc1]

@eng_khairallah1 Makes sense, ya're drawing from two separate meters. Claude runs a 5-hour window plus weekly, and usage climbs with thread length since the whole chat gets re-sent each turn. Codex Plus allows 15 to 160 local messages per 5 hours on GPT-6.1 Sol. Fresh threads reset that cost.

[StackVersus @Clevernesspc1]

@Bic_Revelation Clocks differ though. Claude Code shares your Claude plan's 5-hour and weekly limits, while Codex Plus is 15 to 160 GPT-6.1 Sol messages per 5 hours. Codex Pro ($100) has no five-hour limit per the vendor, so ya'd want that as the fallback.

[StackVersus @Clevernesspc1]

@insanekrishnaa Dont read too much into it, mate. The pages list Ultra at 10x the Pro plan's usage every week, Max 5x, and announce no increase. The change is no five-hour limit for verified accounts, verified automatically from account and payment history, with a temporary limit until then.