OpenAI Renames ChatGPT Pro's 5x and 20x to Standard and More
Recently, OpenAI renamed the original "5x" and "20x" plans of ChatGPT Pro to "Pro Standard" and "Pro More," respectively, and removed the specific descriptions of usage for "5x / 20x Plus" from the pricing page. The official new pricing and limits have not yet been announced.
Community users reported that some interfaces showed a higher tier called "Pro Max," priced at around $500, with A/B testing marketing it as faster Work and Codex. This tier has not appeared in the publicly available official pricing documentation and remains a page experiment and rumor. Starting September 10, OpenAI has suspended new subscriptions and upgrades for the $200 tier (originally Pro 20x), while existing $200 users and $100 tier users are unaffected.
The personal line still maintains a free tier, approximately $8 Go, $20 Plus, and Pro starting at $100. The two Pro tiers were previously divided based on multiples relative to Plus: about $100 for 5x capacity and about $200 for 20x, with higher inference and Codex quotas attached. After the renaming, the homepage no longer reflects the multiple audit, with specific limits remaining in the account usage panel.
Speculation centers on re-layering based on usage intensity and workload, rather than continuing to advertise public multiples. This adjustment comes after the inference costs have sharply increased due to agent conversations, deep research, and coding tasks, and after the top tier expansion has been frozen. On the enterprise side, there are also Business standard seats and higher usage seats, running parallel to the personal Pro renaming.
From a market mechanism perspective, buyers are developers and research users who have hit the Plus limit, while sellers are OpenAI recalibrating compute inventory with tier names. Funds are flowing from personal ARPU to inference and agent conversations. The beneficiaries are those who can channel the paused $200 demand into the new packaging's revenue stream, while the pressure is on heavy users who budgeted based on "20x" but can no longer verify multiples on the pricing page, as well as products like Claude Max at the same price point. The event was driven by the copy changes first, with plans announced later, rather than the already established $500 pricing table.
Source: Public Information
ABAB AI Insight
OpenAI will only split Pro into $100 and $200 in April 2026, benchmarking against Anthropic's Max 5x / 20x. Five months later, the new purchases for the $200 tier are frozen, and the multiples are removed from the homepage, indicating that the public multiples have become a bearish option on GPU bills. The $500 Pro Max, if it only appears on experimental pages, serves to test "is there anyone willing to pay double for faster Work and Codex," rather than immediately changing contracts.
The capital path is to reclaim the quota interpretation rights to the platform. Resources shift from the committed 20x Plus to dynamic quotas within accounts and possibly a new top tier. The motivation is that the marginal costs of agent and coding conversations are rising non-linearly, necessitating the disconnection of users' ability to audit multiples before deciding whether to push $200 users to $500. Strategically, this aligns with API pricing by token: the subscription page sells experience layering, while measurement remains in the background.
Similar paths include cloud vendors changing "unlimited" to fair use, Anthropic using the same 5x / 20x structure, and Google splitting Ultra into two tiers. The industry is entering a control period for usage measurement: the feature lists are already aligned, and the competitive variable has become who can sell between $100 and $500 without explicitly stating multiples.
The structural judgment pertains to the transfer of pricing power. Quota pricing power shifts from the public multiples on the pricing page to account quotas and A/B testing; the $500 is merely a probe to ascertain the demand curve for the top tier. The mechanism is: as the unit session cost rises with agent depth, any multiple written on the official website will be used by users for reconciliation, forcing the platform to first change names and then change bills.
ABAB News · Cognitive Laws
Multiples disappear from the pricing page, and quotas begin to fluctuate
The point where the top tier is paused is the next tier's testing ground
Subscription layering sells intensity, not the multiples in slogans