GTA 6 Costs Different Money On Xbox And PlayStation In Six Markets

Microsoft and Sony do not agree on how to bill Grand Theft Auto VI. Across 32 market pairs read on both storefronts, 26 match to the cent and six diverge, and every one of the six diverges structurally, by currency or by tax, rather than by a few pennies. The Mexican gap is the cleanest: Sony's peso price multiplied by 1.16, which is Mexico's IVA rate, lands on Microsoft's peso price to within three thousandths of a peso, and it does so on two separate editions independently. This desk read both catalogues as primary sources on 2026-11-03, all calls HTTP 200. displaycatalog.mp.microsoft.com/robots.txt returned HTTP 404 at 01:33:04Z, and under the Robots Exclusion Protocol an absent file means unrestricted crawling. store.playstation.com/robots.txt returned HTTP 200 at 143 bytes the same instant, disallowing only /chihiro-api/ and /event/batch with no crawl delay declared, so the product pages were permitted. Both robots files were fetched in a command of their own, before and separate from any content request. One limit belongs at the top rather than buried at the bottom. What follows is a state measured inside the research window, not a proven change inside it. Microsoft's side carries a timestamp and has not moved: the Standard edition record 9P3H4968GRSM reports a LastModifiedDate of 2026-10-23T17:26:34.1280432Z. Sony's side carries no equivalent this desk could read, because the host that would date it returns HTTP 403 and the Internet Archive could not be reached at all. Nobody here can tell you when these six markets began to disagree, only that they disagree now. The six markets where the two stores disagree The comparison is the Standard edition on both stores, Xbox product 9P3H4968GRSM read from displaycatalog.mp.microsoft.com/v7.0/products?bigIds=...&market=<MKT> at 2026-11-03T01:33:20Z, and PlayStation product EP1004-PPSA01547_00-GTAVISTANDARD001 read from store.playstation.com/<locale>/product/... at 01:33:46Z. Prices on the Sony side were taken from the basePrice and basePriceValue fields embedded in each page. MarketPlayStation StoreXbox catalogueNature of the divergence MexicoMXN 1,292.24MXN 1,499.00tax, exactly 16% IVA ArgentinaUS$79.99ARS 113,299.00currency ChileUS$79.99CLP 74,990.00currency ColombiaUS$79.99COP 339,900.00currency United Arab EmiratesAED 299.16USD 79.99currency, reversed Saudi ArabiaSAR 301.56SAR 309.00amount, unexplained Three of the six are one pattern. In Argentina, Chile and Colombia, Sony has dollarised: the page carries basePriceValue 7999 and a currencyCode of USD, so a shopper in Santiago is quoted the same US$79.99 as a shopper in Seattle. Microsoft has not dollarised those three, and bills in pesos. Peru shows the same Sony dollarisation, but the market sweep on the Microsoft side carried no Peruvian record, so there is no pair to compare and Peru is excluded from the count of 32. Official Rockstar Grand Theft Auto VI promotional media. Mexico is exactly IVA, and it holds on both editions Mexico is the one divergence with an arithmetic explanation that closes completely. Sony's Mexican Standard edition price is MXN 1,292.24, carried as basePriceValue 129224. Microsoft's is MXN 1,499.00. The gap is 1,499.00 minus 1,292.24, which is MXN 206.76, and 206.76 divided by 1,292.24 is 16.0001 per cent. Mexico's value added tax is 16 per cent. A single match of that kind can be coincidence, so this desk tested the second edition. Sony's Mexican Ultimate Edition price is MXN 1,619.83. Microsoft's Mexican Ultimate Edition record 9NNZSNHLR63L, read at 2026-11-03T01:34:04Z at HTTP 200, is MXN 1,879.00. Multiply Sony's figure by 1.16 and the result is 1,879.0028. EditionSony, Mexicotimes 1.16Microsoft, MexicoResidual Standard1,292.241,498.99841,499.000.0016 Ultimate1,619.831,879.00281,879.000.0028 Two independently priced products, both landing within three thousandths of a peso of Microsoft's figure. The reading is that Sony displays Mexican prices tax exclusive and Microsoft displays them tax inclusive, which would mean the two stores charge the same money and describe it differently. That reading is an inference from arithmetic and this desk states it as one. Neither product page carries any tax wording whatsoever. A researcher on this cycle initially reported that the PlayStation pages did carry ex-VAT language, and that was wrong: the parser had matched CSS class names such as psw-l-exclude@tablet-s and incidental substrings, not prose. There is no disclaimer to quote, on either side, and anyone reproducing this should not claim there is. The arithmetic is the whole of the evidence, and the arithmetic is very good. Argentina's pesos are hand-set, not converted Microsoft's Argentine prices are worth a second look because they cannot have come from a single exchange rate. Three products, read in the same request at the same instant, imply three different rates against their dollar prices. ProductArgentine priceUS priceImplied ARS per USD Standard, 9P3H4968GRSMARS 113,299USD 79.991,416.4146 Ultimate, 9NNZSNHLR63LARS 141,990USD 99.991,420.0420 Upgrade, 9PN4LLBR8RCHARS 28,691USD 20.001,434.5500 A spread of 18 pesos per dollar across one price list, from one vendor, in one market, at one moment. These are prices chosen to end in tidy local numbers, not prices produced by dividing a dollar figure by a rate. Chile and Colombia show the same character more faintly: 74,990 divided by 79.99 is 937.4922, and 339,900 divided by 79.99 is 4,249.2812, both of which produce round shelf prices rather than round rates. The Gulf pair, one explained and one not The United Arab Emirates inverts the Latin American pattern. Sony bills in dirhams, AED 299.16, and Microsoft bills the UAE in US dollars at USD 79.99, carrying a currencyCode of USD rather than AED. The dirham is pegged at 3.6725 to the dollar, so Sony's AED 299.16 converts to USD 81.4595, which is USD 1.4695 above Microsoft's price, or 1.837 per cent, before any UAE VAT is considered. The same inversion holds on the Ultimate Edition, where Microsoft bills the UAE USD 99.99. Saudi Arabia is the only one of the six this desk cannot explain. Both stores bill in riyals, and the amounts differ. Sony's Standard edition is SAR 301.56 against Microsoft's SAR 309.00, a gap of SAR 7.44, or 2.408 per cent of the Xbox price, with PlayStation the cheaper of the two. If this were a tax presentation difference it would behave like Mexico and scale cleanly, and it does not. On the Ultimate Edition, Sony is SAR 376.96 against Microsoft's SAR 389.00, a gap of SAR 12.04, or 3.095 per cent. The gap is not constant as a percentage and not constant in absolute terms, so it is not a single tax or fee applied consistently. Dividing Microsoft's SAR 309.00 by 1.15, for the Saudi 15 per cent VAT rate, gives 268.70, which matches neither figure. This desk reports the discrepancy and offers no mechanism for it. What matches, and it is most of it Six divergences out of 32 pairs means 26 markets where the two storefronts agree exactly, including the United States, Canada, the United Kingdom, Ireland, Brazil, India, South Africa, Turkey, South Korea, Hong Kong, Taiwan, Singapore and fourteen further European markets. The headline $79.99 and $99.99 prices are old news here and are not the story. Two further checks came back clean. Edition arithmetic is perfect in all 40 Microsoft markets swept: Standard plus Ultimate Edition Upgrade equals Ultimate Edition, with a delta of 0.00 in 40 of 40. In the United States that is 79.99 plus 20.00 equals 99.99; in Argentina 113,299 plus 28,691 equals 141,990; in Japan 9,800 plus 2,480 equals 12,280; in the United Kingdom 69.99 plus 20.00 equals 89.99. No market lets a buyer assemble the Ultimate Edition for less than its sticker price, which is worth knowing because the question gets asked every week. Unlock times also agree across the two stores in 35 of 35 markets, to the second. Sony's releaseDate equals Microsoft's OriginalReleaseDate in every one of the 32 pairs plus Australia, New Zealand and Japan, and every value is local midnight on 19 November, correct for daylight saving. The spread runs from New Zealand at 2026-11-18T11:00:00Z to Mexico at 2026-11-19T06:00:00Z, a rolling launch of 19 hours. This desk has published that 19 hour framing before, so the new part is only that both storefronts now corroborate it. One oddity surfaced in the same data: Taiwan's pre-order window opened at 2026-09-16T14:00:00Z, roughly twelve weeks after every other market, which opened on 24 or 25 June. All 40 windows close at local 23:59:59 on 18 November. Still no Steam listing, while GTA+ is already a Steam product Sixteen days from launch there is no Grand Theft Auto VI application in Steam's public store search. Read at 2026-11-03T01:34:26Z at HTTP 200, the query term=grand+theft+auto+vi returns total: 1, and the single hit is app 1546990, Grand Theft Auto: Vice City, The Definitive Edition, which matches on a substring rather than being the game. A control query for term=cyberpunk returned total: 10 including Cyberpunk 2077 at app 1091500, so the endpoint works and the absence is a real absence rather than a broken request. store.steampowered.com/robots.txt returned HTTP 200 at 303 bytes and does not disallow /api/; api.steampowered.com disallows everything and was left alone. The same search surfaces something more interesting. GTA+, the subscription that every Grand Theft Auto VI pre-order bonus is denominated in, is already a live Steam product at app 3482260, priced USD 7.99. The game has no Steam presence and its subscription service does. Sidebar: the record nobody can buy now holds 19,828 ratings Batch 70 reported that Microsoft's achievement, presence and cloud save flags sit on a carrier record for Grand Theft Auto VI that no customer can transact against. That record has since accumulated a remarkable quantity of ratings while reporting zero purchases and zero playtime. Read at 2026-11-03T01:34:22Z, HTTP 200, 178,376 bytes, market US. RecordAll-time ratingsLast 30 daysLast 7 daysPlayCountPurchaseCount 9NL3WWNZLZZN carrier, not purchasable19,8285,3772,07100 9P3H4968GRSM Standard, $79.99101000 9NNZSNHLR63L Ultimate3011200 9PN4LLBR8RCH Upgrade103200 The three records customers can actually buy hold 50 ratings between them. The one they cannot buy holds 19,828, which is 396.6 times as many. The baseline for comparison is this desk's own batch 41, whose fetch was timestamped 2026-08-31T22:55Z and recorded 11,965 all-time ratings with 931 in the preceding seven days. That is 7,863 new ratings over 64 days, an average of 122.86 a day. The current seven day pace is 2,071 divided by 7, or 295.8571 a day, against 133.0 a day at the August reading, a multiple of 2.2245. Against the preceding 23 days of the 30 day window, which carried 3,306 ratings at 143.7391 a day, the current week runs at 2.0583 times the rate. What this is not: play data. PlayCount is 0 in every window on every record. Nobody has played the game. A store rating is a listing figure that Microsoft's catalogue exposes, and the thing being rated here is a page. Sidebar: a second Vanguard adviser crossed 5% of Take-Two, and it breaks our own table The only filing of any kind on Take-Two's EDGAR file dated 28 October or later is a Schedule 13G, and it rewrites an ownership table this desk published in August. www.sec.gov/robots.txt returned HTTP 200 at 2,736 bytes when fetched at 2026-11-03T01:26:25Z, carrying a single wildcard group with an explicit Allow: /Archives/edgar/data, no crawl delay and no ClaudeBot rule. All SEC requests declared the user agent "6Charts News Desk [email protected]". Accession 0002100121-26-001132 was accepted at 2026-10-30T18:13:34.000Z, which the index page renders as 14:13:34 Eastern. The filer is Vanguard Portfolio Management LLC, CIK 0002100121, reporting 9,601,771 shares of Take-Two, or 5.13 per cent, with an event date of 09/30/2026, filed under Rule 13d-1(b) as a passive institutional holder, type IA, CUSIP 874054109, SEC file number 005-52733. Sole voting power is 46,943 against sole dispositive power of 9,601,771. It is signed by My Trieu-Gatt, Authorized Signatory and Head of Global Fund Administration. The Vanguard reporting entity for Take-Two has been split in three stages, and this desk read all three primary documents. FilerEvent dateSharesPer cent The Vanguard Group03/13/202600 Vanguard Capital Management LLC03/31/202613,131,9087.09 Vanguard Portfolio Management LLC09/30/20269,601,7715.13 Batch 36, published 2026-08-25, carried Take-Two's proxy ownership table with a single line reading "The Vanguard Group, 13,131,908, 7.02%" and described Vanguard as the largest holder, marginally ahead of BlackRock. That share count is exactly Vanguard Capital Management's 13G figure, so the proxy line was built from one filing. As of 30 October there are two Vanguard 13Gs on file rather than one, and the disclosed total is 13,131,908 plus 9,601,771, or 22,733,679 shares. The two stated percentages sum to 12.22 per cent, though they carry different as-of dates and therefore different denominators; on a 186.98 million share count the combined holding is 12.158 per cent, and 9,601,771 against that count is 5.135 per cent, which ties the filing's own 5.13. At the $192.10 close the new block is worth $1,844,500,209 and the combined holding $4,367,139,736. The filing landed 20 days before launch. The caveat is as important as the finding. A Schedule 13G cannot tell you whether this represents net buying or an internal reallocation between Vanguard advisory units, and the filing's own comments field explicitly disaggregates Portfolio Management from other Vanguard affiliates. The defensible claim is that the disclosed record moved from one 7.02 per cent line to two filings totalling roughly 12.2 per cent, and that a Vanguard entity reported crossing 5 per cent of Take-Two for the first time as of 30 September. Nothing stronger than that is supported, and this desk asserts nothing stronger. Take-Two itself has filed nothing in the window. Its most recent own filings are a Form 4 on 2026-10-16 and two Rule 144 notices on 2026-10-15. The control for that negative is that the same API call returned those known filings, so the endpoint and the CIK are good and the silence is the record rather than a broken query. The company reports second quarter results after the close on Monday 9 November, six days from today, with a call at 4:30pm Eastern. Three things circulating that this desk will not print as fact A Nike collaboration is the liveliest of them. Three Air Force 1 colourways are reported at $130 each for 19 November, carrying style codes JJ7602-100, JJ7604-001 and JJ7603-900, under a Cactus Jack and Rockstar Games billing. The trail runs to a Canadian retailer's release calendar labelled "Nike x Cactus Jack x RS" and a Romanian sneaker leak account, then to sneaker media the same afternoon. Three outlets publish the identical code triplet, and that is consistency rather than independence, because all three plausibly descend from the same two origin artefacts. The retailer's calendar itself could not be inspected, because that host was unreachable through this desk's proxy, so the first link in the chain is unverified here. Against it sits one primary negative: nike.com/launch, fetched at 2026-11-03T01:18Z at HTTP 200, lists none of it, while rendering eleven real products and a dated SNKRS editorial slot, so the page was working. One point in the story's favour deserves saying plainly, because almost nothing else this cycle clears the same bar: 2026 Nike style codes cannot have come from the 2022 Rockstar breach, so the usual contamination route is structurally impossible here. For scale, $130 is $50.01 more than the Standard edition of the game, and three pairs at $390 is 4.88 times the game's price. Credible, unconfirmed, and not a 6Charts report. The 210 GB download size remains one X account's scrape, dated 27 October and therefore outside this window, relayed onward by several outlets. No first party surface carries it. Microsoft's catalogue exposes only a 344,731,648 byte pre-order stub, which is 328.76 MiB, with MaxInstallSizeInBytes at 0, and Sony's web product pages contain no fileSize field at all. The host that would carry the real figure returns HTTP 403 to this desk. The claim of 201 main story missions came from an X account with five posts in total and no cited source, on 28 October, which is also outside the window. It can be tested, and it fails. On 26 October the same account claimed the game would be 230 GB, against the 210 GB figure everyone else is repeating, an overstatement of 20 GB or 9.52 per cent, and when challenged it moved the claim to 230 GB after a day one patch. An account whose only checkable prior number missed and was then retrofitted has a negative track record rather than an unproven one. Limits, and what this desk did not do rockstargames.com carries User-agent: ClaudeBot with Disallow: / site-wide. This desk did not read that host this cycle, not even its robots file, and retrieved no Rockstar media assets. The standing brief names rockstargames.com/VI/media/screenshots as the image source, so the image sourcer declined that path and selected from assets already held in this repository. x.com and reddit.com disallow all automated clients, so no primary post text was read anywhere this cycle and every claim above attributed to a post there is second hand through search snippets. No agent substituted or disguised a user agent at any point. Permission could not be established for the following, all of which were left alone: web.np.playstation.com, the GraphQL host that carries PlayStation install sizes, HTTP 403; web.archive.org, connection reset, which is the direct reason the pricing finding cannot be dated on Sony's side; investing.com and Take-Two's own investor newsroom, both 403; classification.gov.au, an HTTP/2 error and then a timeout; pegi.info, an HTTP/2 error; sneakernews.com, 403; forbes.com, a 302 with an empty body; the Canadian retailer host behind the Nike lead, a proxy 502; the USPTO search API, 405 to POST; kalshi.com, where the ClaudeBot group permits only a handful of documentation paths and not market pages; psdeals.net, dexerto.com, complex.com and bloomberg.com, all carrying ClaudeBot Disallow: /; trends.google.com, whose data endpoints are disallowed; and api.steampowered.com, a wildcard disallow. ESRB's rating detail pages are disallowed to this agent and the host publishes Crawl-delay: 20, so no ratings claim appears today. Nobody at Rockstar Games, Take-Two, Sony, Microsoft, Nike or Vanguard was contacted, and nothing above establishes intent. This desk prints its own faults, and this cycle produced several worth recording. The first is the editor's, and it changed the headline. The researcher who ran the price sweep summarised it as 28 of 32 markets matching with four divergences, but its own market table lists 26 matching markets and six divergent ones, which sums correctly to 32. Six is right. The editor verified all six divergences first hand, plus a United States control, before correcting the figure. A summary line that contradicts the table beneath it is the easiest error in this trade to pass along. Three more came from the platform researcher, all disclosed before filing. It nearly reported that the PlayStation record had lost Remote Play support, having read only the first compatibilityNoticesByPlatform block in the page, a trimmed copy embedded by one micro-frontend; the full payload later in the same document carries REMOTE_PLAY_SUPPORTED as true, and nothing had been removed. It reported fake evidence of a tax disclaimer, described above. And its preload check was language blind, searching for the English string "7-Day Pre-load" and returning false for all twenty non-English locales, which would have supported a false story that preload is not offered outside English markets; the Japanese page carries the equivalent string eight times, and this desk makes no claim about which locales list preload. Two came from the legal and corporate researcher. It nearly cleared Bloomberg by accident: its first robots parser printed only the opening 25 lines and only the wildcard group, which made the host look open, while the ClaudeBot Disallow: / sits at line 95 of a 318 line file. It caught this before fetching any Bloomberg content, and the structural lesson is that a truncated robots check is a broken robots check. It also measured its own tooling rather than the record, running novelty searches with a 250 character window in a mode where the wildcard does not match newlines, which returned empty against multi-line seed files and briefly made an already published story look new. Two came from the community researcher. It built an attention metrics line on an assumption that failed: YouTube's robots file permits /watch, but the channel page is rendered in JavaScript and two /watch requests returned 302 redirects to a CAPTCHA interstitial, which it did not follow, retry or work around. The consequence is printed rather than hidden: no primary view count, like count or comment count exists anywhere in this cycle, and any trailer view figure in circulation today is secondary. It also found three prediction market entries whose structured data conflicted with their displayed prices and passed the conflict through rather than choosing the better looking number, so those three are omitted entirely. Finally, the image sourcer's count of articles since its chosen image was last used was off by one, 395 against a true 394, caught by the editor recomputing per file; the selection stands by a wide margin either way. Further coverage from this desk sits on the news index, and servers preparing for the 19 November launch are catalogued on the server list.