FOURWEEKMBA TRACKER ยท FROM THE NEWSROOM ยท UPDATED 2 NOVEMBER 2026
Free to share with a link to this page. Figures from Anthropic’s S-1 revenue table; multiples and shares are our arithmetic.
The short version: the flat-rate AI subscription is turning into a meter. In 2023 the market had one price point, $20 a month. Today the six companies on this page sell 25 paid consumer and individual plans from $4.99 to $500 a month, and more of them count usage in tokens, compute or credits rather than messages. GitHub Copilot has billed by credits since 1 June 2026, Gemini moved to compute-based limits on 17 May 2026, Grok shares one usage pool across products since 12 October, and OpenAI says ChatGPT and Codex will share one token-based allowance from December.
Anthropic’s S-1 shows why the plans matter to the labs: subscriptionrevenue was $3,222 million in the six months to 30 June 2026, 17.9 times the year before, while API revenue grew 13.3 times, so subscriptions rose from 15.5% to 19.8% of revenue (our arithmetic; S-1). On 1 November Elon Musk wrote that “the cost of serving Grok Bot was unsustainably high relative to revenue” (post). This page tracks every plan, its limits and each dated change. US prices, read on 2 November 2026.
32 plans from 6 companies. “Metered” means usage is counted in tokens, compute or credits, or extra usage can be bought, beyond the flat price. Numbers in brackets link to the source for that row. Business and enterprise seats are not listed.
Plan
US price
Usage limits
Metered?
Last change and sources
OpenAI
ChatGPT Free
$0
Unlimited text chats subject to abuse guardrails; limited uploads, images, deep research, Codex
No change found
Ads tested on Free and Go since 9 Feb 2026 (US) [1][2]
ChatGPT Go
$8/mo may include ads
More messages, uploads and images than Free
No change found
Rolled out worldwide 16 Jan 2026 at $8 in the US [1][2]
ChatGPT Plus
$20/mo
Five-hour and weekly windows in Work and Codex; Codex usage rebalanced across the week (9 Apr 2026)
Yes, from December: one token-based allowance for ChatGPT and Codex; credits and paid instant resets at the limit
31 Oct 2026: shared token allowance announced for December [1][2][3][4]
ChatGPT Pro 100
$100/mo
No five-hour limit in Work and Codex; weekly allowance; doubled Work and Codex allowance 29 Oct 2026
Yes, from December (shared token allowance); $100/month API credits
29 Oct 2026: Work and Codex allowance doubled [1][2][3][4]
ChatGPT Pro 200
$200/mo
Allowance cut announced 29 Sep 2026, cancelled 29 Oct 2026; lower allowance doubled to match the higher one
Yes, from December (shared token allowance); $200/month API credits
The entry price has not risen. ChatGPT Plus is still $20 a month, as at its launch in February 2023, and Google cut AI Plus from $7.99 at its US launch in January 2026 to $4.99 (reported by Engadget on 8 June; Google’s US page lists $4.99). What changed is the top of the ladder: OpenAI sells Pro at $100, $200 and $500, Anthropic, xAI and Perplexity top out at $200, Google at $199.99 and Microsoft at $99.99 for Microsoft 365 Pro. Google also cut its top Ultra plan from $250 to $200 in May 2026 while adding a $100 tier.
What changed: limits moved to meters
The second change is how usage is counted. Five-hour windows with a weekly cap are now the norm at Anthropic (weekly limits since 28 August 2025), Google (since 17 May 2026) and OpenAI’s Work and Codex. GitHub said its old model was “no longer sustainable” because “a quick chat question and a multi-hour autonomous coding session can cost the user the same amount” (GitHub, 27 April 2026). OpenAI says that from December, Plus, Pro and Business usage “will be measured in tokens” and that at the limit users can “wait for it to reset, upgrade your plan, or buy credits” (OpenAI Help).
Not every move tightened. Anthropic removed the 5-hour limit for verified Ultra accounts on 30 October and said weekly limits are unchanged; OpenAI cancelled a planned Pro 200 cut and doubled Work and Codex allowances on 29 October, and added monthly API credits equal to each Pro tier’s price. The direction is toward measured usage, not simply less of it.
12 Oct 2026 ยท xAI: Grok, Cursor and X Premium unified into one shared usage pool, $8 to $200 (@SpaceXSI on X, 12 Oct 2026, via our report (via our newsroom))tighter / metered
For subscribers: heavy agentic use (coding agents, Grok Bot, ChatGPT Work) is what the meters now price. Light chat use is mostly unchanged; OpenAI says Instant chat “will not draw from your usage”.
For investors in the labs: Anthropic’s S-1 lists among its risks that “we have also changed, and may in the future change, usage limits, rate limits, or model and product capabilities, which could adversely affect our reputation and the perceived value of our offerings.”
For software buyers: Microsoft chief executive Satya Nadella said on 28 October that “our business model is expanding from per-seat subscriptions to include consumption”, with Copilot customers on usage-based billing up over 75% in the quarter (our report).
The Business Engineer framework The Subsidized AGI Economy argues that a plan priced below what its heaviest users consume is doing strategic work. It is the author’s argument and predates this week’s changes.
What is not established: no company on this page says whether its subscriptions lose money. Musk gave no figures for Grok Bot’s cost or revenue. SemiAnalysis’s estimate that Claude plans can take over 40% of Anthropic’s inference compute is its own rough estimate, not Anthropic’s (our report). OpenAI has not published December allowance sizes.
FAQ
What is the “AI subscription subsidy”?
The gap between a flat monthly price and what a heavy user’s usage would cost at API list prices. A guest on a SemiAnalysis clip on 7 October said “you can pay $200 for a Claude plan and get $12,000 of Opus 5.5 tokens at API pricing”, 60 times the price by our arithmetic, on that guest’s workload. Meters and weekly caps narrow that gap for the heaviest users.
Are AI subscriptions getting more expensive?
The entry plans are not: ChatGPT Plus and Claude Pro are still $20, ChatGPT Go is $8 and Google AI Plus is $4.99. The top tiers are: they now run from $99.99 to $500 a month, against $20 for each company’s first paid plan.
Which AI plans are usage-based now?
GitHub Copilot bills every plan by AI credits since 1 June 2026. Gemini uses compute-based limits that refresh every 5 hours up to a weekly limit, extendable with AI credits. Grok, Cursor and X share one usage pool. Perplexity Pro and Max run on credits (100 credits = $1). OpenAI’s Plus and Pro move to one token-based allowance in December. Claude plans keep 5-hour and weekly limits, with API credits on Max and Ultra.
Did anyone loosen limits?
Yes. Anthropic removed the 5-hour limit for verified Ultra accounts on 30 October 2026. OpenAI cancelled its Pro 200 allowance cut and doubled Pro 100 and Pro 500 Work and Codex allowances on 29 October. Google cut its top Ultra plan from $250 to $200 in May 2026 and AI Plus to $4.99 in June.
Can I compare limits across companies?
Not directly. Each company measures in its own unit: multiples of Pro (Anthropic, Google), multiples of Plus (xAI), dollars of credits (GitHub, Perplexity) or tokens (OpenAI from December). The table quotes each company’s own wording.
How often is this page updated?
Whenever our newsroom logs a new plan change, and at least weekly against the companies’ pricing pages. Each row carries its source and date.
2026-10-28 Microsoft (Nadella, official prepared remarks): "Our business model is expanding from per-seat subscriptions to include consumption"; Copilot usage-billing customers +75% q/q.
2026-10-30 Anthropic (@claudeai 2116220869479428500): Max 5x->Max $100, Max 20x->Ultra $200; 5-hour limit removed for verified Ultra; "Weekly limits remain unchanged."
2026-10-31 OpenAI help center 20001625: from December Plus/Pro/Business allowance shared across ChatGPT and Codex, measured in tokens; at limit: wait, upgrade or buy credits.
2026-11-01 20:34Z Elon Musk (2116992258520064397): "The cost of serving Grok Bot was unsustainably high relative to revenue" (~5.77M views 2 Nov).
2026-11-01 21:48Z Amjad Masad (2117010858845089805): "If you're enjoying a subsidized AI subscription get ready to get rug pulled." (earlier 2065180078439547275, 11 Jun 2026: "The rug pull is coming… sooner or later.")
2026-10-07 SemiAnalysis post 2107950798638354671 (guest quote): "$200 for a Claude plan and get $12,000 of Opus 5.5 tokens at API pricing" (60x, our arithmetic). 5 Oct article: subs ~10% revenue, "over 40% of inference compute" (estimate).
How this page stays current
Built by FourWeekMBA from each company’s pricing pages, help centres, changelogs and announcements, read on 2 November 2026 (US view), and from our newsroom’s sourced log on the end of the AI subscription subsidy. Items marked “reported” come from the named outlet and are not yet confirmed on the company’s own page. Counts, multiples and shares are our arithmetic. When the newsroom logs a new change, the log above updates and the table, timeline and charts are rebuilt.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
