I’m a ChatGPT Business user and my usage pattern is heavily Codex-first.
In practice, I use ChatGPT.com very little, while most of my actual professional workload happens in Codex. The current packaging feels inefficient: I get a large weekly ChatGPT.com message allowance that I barely consume, but the Codex limits are the bottleneck in my daily work.
This creates a mismatch between what I pay for and what I actually use.
A better structure would be a Codex-first or Codex-only plan/seat:
pricing aligned with developer workflows rather than general chat usage.
For users like me, buying additional ChatGPT seats or higher general-purpose plans is not the right solution, because the scarce resource is not ChatGPT.com usage. The scarce resource is Codex capacity.
The issue is not that I expect unlimited usage for free. The issue is that the current bundle allocates too much value to a product surface I do not use, and too little capacity to the product surface I use all day for work.
A Codex-first plan would be much more useful for solo developers, small technical teams, and Business users who use OpenAI primarily as an engineering tool rather than as a general chat assistant.
ChatGPT Business supports Codex seats with usage-based pricing. There is no fixed monthly seat price, no required minimum number of Codex seats, and Codex-only access does not include access to the ChatGPT workspace.
Thanks for the pointer, but the Codex seat with usage-based pricing is essentially the same model as using the API directly — you pay per token consumed, with no predictable monthly cap. That’s not really a subscription; it’s metered billing wrapped in a Business interface.
The core ask here is different: a flat monthly plan oriented around coding workflows, similar to what Kimi or DeepSeek already offer with their developer/coding-focused subscriptions. Those give you a fixed quota you can rely on every month, which is crucial when you’re using a coding tool 8 hours a day in a professional environment.
For that kind of usage pattern, usage-based pricing introduces unpredictable costs and doesn’t give you the peace of mind of a capped seat. At that point, going straight to the API is actually more flexible and transparent — you at least have full control and no Business plan overhead.
A proper Codex subscription tier (flat fee, high coding capacity, minimal ChatGPT.com allowance) would be a much better fit for developers and teams using Codex as their primary work tool.
Or, hey - here’s a crazy thought: just allow Business users to upgrade individual high-usage seats to the Pro-tier options. We’re here and ready to give you more money. We’ve been a loyal Business/Teams customer for as long as the plan ever existed. Yet we’re relegated to the same usage limits as Plus, we have no need for the fancy enterprise SSO-esque add-ons, we DO need Pro quota caps, and we’re forced NOT to upgrade because the only way to do so is either to pay for our current Business seats AND separate disconnected Pro seats for the same human users who have no means to apply multiple Codex accounts on the same computer without constantly de-authing and re-authing between these accounts, OR, delete our Business account and seats and lose over a year of saved memories and files and dozens of projects and thousands of past chat sessions. WHAT are you doing to your loyal Business customers, OpenAI? You seem to treat your FREE users better than us, and that’s NOT just lately; we literally only got the annual Business plan in the first place because it was announced that memory would be made available only for that level of user… then, the memory feature turned out to have the storage capacity of a TI-83 at launch time, and long before that got any better, free and Plus users were given access to the feature you advertised to us as “exclusive” at the outset of our Business relationship.
C’mon. I mean… come on. That’s about as weak as sauce can get.
And also, just for the record: no, for a small business with 2 employees the statement, “All you have to do is pay like a firehose through the nose by using the API instead of the subscription plan that every Average Joe is offered but is denied to the higher-price Business plan customers” is nothing short of objectively an abject and offensive insult.
I definitely think a per seat Business Pro plan would be super helpful and we have team members would could use it right away.
I should mention that a “Codex first” plan makes no sense financially for OpenAI, so the Business Pro is a better ask since it would mirror what individual Pro users get.
I think “Business Pro” would help some teams, but it does not fully solve the affordability problem for small teams, freelancers, and individual developers using Codex as their primary work tool.
The gap is not only between Business and Pro. The gap is between predictable subscription usage and metered Codex usage.
A Pro-like Business seat may be too expensive for many small businesses or solo developers. What I’m asking for is not Pro-level quota at a lower price. I’m asking for a developer-oriented Business seat with a different allocation of the same kind of subscription value:
pricing accessible to small teams and freelancers.
This could sit below Pro, so it would not need to offer Pro-level capacity. But it would be much better aligned with how many developers actually use the product.
Right now, the choices are awkward:
pay for Business seats where much of the included ChatGPT.com capacity goes unused;
use Codex seats with usage-based billing, which is less predictable;
buy separate Pro accounts, which fragments workspace, memory, files, and projects;
move to Enterprise, which is not realistic for many small teams.
There should be a middle tier: a Codex-focused Business Developer seat.
The point is not unlimited usage. The point is predictable, affordable, business-safe coding capacity.