OpenAI Added a $100 ChatGPT Pro Tier for Heavier Codex Use

OpenAI's new subscription change looks small if you read it as another pricing page tweak. It is more revealing if you read it as a demand signal.

The company has added a new $100 ChatGPT Pro plan between Plus and the existing $200 Pro tier. In OpenAI's plan documentation, the company says the new tier is built for users who need advanced tools throughout the week, offers 5x higher limits than Plus, and gives 10x Codex usage versus Plus for a limited time. The existing $200 Pro plan remains in place as the heaviest-usage option, now positioned at 20x usage versus Plus.

That structure matters because it turns a fuzzy market trend into a price ladder. OpenAI is effectively saying there is now a meaningful group of users whose behavior sits between casual experimentation and nonstop professional dependence. They use Codex and other advanced tools enough to pay more than Plus, but not enough to justify the highest tier.

Pricing changes often reveal more than product launches do. Companies can demo future ambitions all day. Pricing is where they admit what sort of customer behavior they think is really happening. A new $100 middle tier says OpenAI believes sustained coding-agent use has grown enough to deserve its own monetization lane.

That fits the rest of the company's recent signals. Codex is being pushed harder across ChatGPT and developer workflows, and our earlier look at OpenAI's Codex plugin packaging work pointed to the same direction from the product side. OpenAI is not treating coding agents like a side experiment anymore. It is building a fuller usage economy around them.

The new pricing also clarifies what OpenAI thinks the core segmentation is. Plus at $20 is for lighter use. The new $100 Pro tier is for real projects and regular use of advanced tools. The $200 Pro tier is for the people who want to run harder workflows continuously and across parallel projects. That is a much more operational product story than the old consumer-style subscription framing.

The significance is not that $100 is a magical number. The significance is that OpenAI thinks there are enough people doing serious work with Codex that a new step between hobbyist and power user now makes commercial sense.

Why Codex Is Driving the Pricing Story

OpenAI's own language makes Codex central to the change. The help-center page says the $100 Pro tier is aimed at users who need advanced tools and models throughout the week, with 10x Codex usage versus Plus for a limited time. That tells you where the pressure is coming from. Users are not only chatting more. They are running longer, heavier coding workflows that need more usage headroom.

That changes how you should think about coding-agent adoption. A year ago, many teams were still trying AI coding tools for light edits, suggestions, and one-off assistance. Those patterns fit comfortably inside a low-cost consumer plan. Once users start relying on coding agents for sustained sessions, cloud delegation, larger repo work, and repeated task loops, the old pricing tiers stop matching the behavior.

OpenAI is solving that mismatch by separating "I use this sometimes" from "I work this way regularly." That is a healthy sign for the category because it means usage is becoming legible enough to price directly. Markets mature when vendors can identify stable patterns and build offers around them.

The Codex-specific language matters for another reason. Coding agents are unusually expensive to serve compared with simple chat prompts. They read more context, keep longer sessions alive, run more reasoning steps, and often trigger tool use or background execution. A company that sees heavy Codex demand has to decide whether to cap it aggressively, hide it inside a broad premium plan, or create a clearer pricing rung. OpenAI has chosen the third path.

That is also a hedge against customer frustration. If too many serious users crowd into the $20 plan and constantly hit limits, the product feels stingy. If the only upgrade path jumps straight to $200, many users will see the ladder as too steep. A $100 middle tier gives OpenAI a way to catch the users who want more room without making them feel forced into the top shelf.

In plain terms, this is not just about charging more money. It is about matching price to intensity of use so the product keeps feeling usable as the workload changes.

What the New Tier Says About the Market

The clearest reading is that coding-agent usage is fragmenting into tiers faster than basic AI chat. Plenty of people still want occasional help with drafting, search, or light coding. Another group now wants a system that can stay with them through real development work, which means longer sessions, more retries, more repo context, and less tolerance for hitting a wall halfway through a task.

That matters for competitors too. If OpenAI sees enough demand to support a middle Pro tier, then other vendors will likely face the same segmentation pressure. The AI subscription market may end up looking less like streaming video and more like cloud software, where heavier operational use naturally creates more pricing layers.

There is also a business-model point here. OpenAI has spent years balancing broad consumer appeal with expensive frontier-model usage. Coding agents sharpen that tradeoff because they are both sticky and costly. If Codex users become some of the most valuable customers on the platform, then pricing has to acknowledge that without scaring away the broader base.

The $100 tier is a compromise. It lets OpenAI capture more revenue from committed users while still keeping an entry point at $20 and a top tier at $200. It also helps the company message that "Pro" is no longer one undifferentiated bucket. There are now degrees of professional AI usage, and OpenAI wants to sell into each of them.

That tells enterprise buyers something too. Consumer and self-serve pricing often previews what vendors are learning about workload intensity before those lessons show up in larger contracts. If a vendor is slicing self-serve plans around coding-agent depth, it probably means the underlying cost and usage patterns are becoming clearer across the whole business.

How Teams Should Read This Change

The first mistake is to treat the new tier as only a personal-budget question. If you run a team, the more useful question is which workflows are moving from occasional AI assistance into regular agent dependence. Pricing changes like this can expose where your engineers are crossing that line before you have formalized it internally.

The second mistake is to compare plans only on sticker price. A better comparison is cost per week of useful work. If a developer keeps bouncing off Plus limits during important Codex sessions, the real cost includes interruption, context rebuilding, and time lost waiting for the next window. In some cases a higher tier can be cheaper than unreliable access.

The third mistake is to assume the market has settled. OpenAI is still experimenting with temporary multipliers and adjusted limits. The current 10x Codex boost on the $100 tier runs through May 31, which means the company is still shaping where sustainable demand lands. Buyers should read this as a live pricing system, not a final steady state.

Still, the direction is clear. Coding-agent work is becoming important enough to reorganize pricing around it. That is a bigger development than one more subscription option. It means AI vendors are no longer pricing only for curiosity and broad productivity. They are pricing for sustained agent use that looks closer to real software work.

OpenAI's new $100 plan is therefore a market message disguised as a billing update. The company thinks Codex use is heavy enough, valuable enough, and distinct enough to deserve its own tier. Whether competitors follow quickly or reluctantly, that is the signal worth watching.