OpenAI is making strikes to attempt to courtroom extra builders and vibe coders (those that construct software program utilizing AI fashions and pure language) away from rivals like Anthropic.
Immediately, the agency arguably most synonymous with the generative AI increase introduced it would start providing a brand new, extra mid-range subscription tier — a $100 ChatGPT Professional plan — which joins its free, Go ($8 month-to-month), Plus ($20 month-to-month) and current Professional ($200 month-to-month) plans for people utilizing ChatGPT and associated OpenAI merchandise.
OpenAI additionally at the moment affords Edu, Enterprise ($25 per consumer month-to-month, previously often known as Workforce) and Enterprise (variably priced) plans for organizations in stated sectors.
Why supply a $100 month-to-month ChatGPT Professional plan?
So why introduce a brand new $100 ChatGPT Professional plan, then?
The large promoting level from OpenAI is that the brand new plan affords 5 instances higher utilization limits on Codex, the corporate's agentic vibe coding utility/harness (the title is shared by each, in addition to a lineup of coding-specific language fashions), than the prevailing, $20 month-to-month Plus plan, which appears honest given the mathematics ($20×5=$100).
As OpenAI co-founder and CEO Sam Altman wrote in a submit on X: "It is vitally good to see Codex getting a lot love. We’re launching a $100 ChatGPT Professional tier by very fashionable demand."
Nevertheless, alongside this, OpenAI's official firm account on X famous that "we’re rebalancing Codex utilization in [ChatGPT] Plus to assist extra classes all through the week, fairly than longer classes in a single day."
That sounds rather a lot like OpenAI can also be concurrently lowering how a lot ChatGPT Plus customers can use its Codex harness and utility per day.
What are the brand new utilization limits for the brand new $100 ChatGPT Professional plan vs. the $20 Plus?
So, what are the present limits on the $20 Plus plan? The brand new Professional plan provides you 5X higher than…what?
Seems, that is trickier than you'd suppose to calculate, as a result of it really varies relying on which underlying AI mannequin you might be utilizing to energy the Codex utility or harness, and whether or not you might be engaged on code saved within the cloud or regionally in your machine or servers.
OpenAI’s Developer web site underwent a number of updates at the moment, so we've solely mirrored the most recent pricing construction and choices under as of Thursday, April at 10:45 pm ET. It notes that for particular person customers, Codex utilization is categorized by “Native Messages” (duties run on the consumer’s machine) and “Cloud Duties” (duties run on OpenAI’s infrastructure), and people limits share a five-hour rolling window.
It additionally says further weekly limits could apply. The present Codex pricing web page now reveals decrease displayed utilization ranges than the older model, and it measures Code Opinions in a five-hour window fairly than per week. For Professional 5x particularly, OpenAI says the at the moment proven limits embody a short lived 2x utilization increase that ends Could 31, 2026.
ChatGPT Plus ($20/month)
-
GPT-5.4: 20–100 native messages each 5 hours.
-
GPT-5.4-mini: 60–350 native messages each 5 hours.
-
GPT-5.3-Codex: 30–150 native messages and 10–60 cloud duties each 5 hours.
-
Code Opinions: 20–50 each 5 hours.
ChatGPT Professional 5x ($100/month)
-
GPT-5.4: 200–1,000 native messages each 5 hours.
-
GPT-5.4-mini: 600–3,500 native messages each 5 hours.
-
GPT-5.3-Codex: 300–1,500 native messages and 100–600 cloud duties each 5 hours.
-
Code Opinions: 200–500 each 5 hours.
Notice: The bounds proven for Professional 5x embody a short lived 2x utilization increase that ends Could 31, 2026.
ChatGPT Professional 20x ($200/month)
-
GPT-5.4: 400–2,000 native messages each 5 hours.
-
GPT-5.4-mini: 1,200–7,000 native messages each 5 hours.
-
GPT-5.3-Codex: 600–3,000 native messages and 200–1,200 cloud duties each 5 hours.
-
Code Opinions: 400–1,000 each 5 hours.
-
Unique entry: Contains GPT-5.3-Codex-Spark in analysis preview for ChatGPT Professional customers solely. OpenAI says it has its personal separate utilization restrict, which can alter primarily based on demand.
And as OpenAI's Assist documentation states:
"The variety of Codex messages you possibly can ship inside these limits varies primarily based on the dimensions and complexity of your coding duties, and the place you execute duties. Small scripts or easy features could solely eat a fraction of your allowance, whereas bigger codebases, lengthy operating duties, or prolonged classes that require Codex to carry extra context will use considerably extra per message."
The bigger strategic implications and context
OpenAI’s sudden transfer towards the $100 value level and expanded agentic capability comes amid the unprecedented monetary ascent of its chief rival, Anthropic.
Simply days in the past, Anthropic revealed its annualized run-rate income (ARR) has topped $30 billion, surpassing OpenAI's final reported ARR of roughly $24–$25 billion.
This development has been fueled by the large adoption of Claude Code and Claude Cowork, merchandise which have set the benchmark for enterprise-grade autonomous coding.
The aggressive friction intensified on April 4, 2026, when Anthropic formally blocked Claude subscriptions from getting used to supply the intelligence for third-party agentic AI harnesses like OpenClaw.
To be clear, Anthropic Claude fashions themselves can nonetheless be used with OpenClaw, customers simply should now pay for entry to Claude fashions via Anthropic's utility programming interface (API) or additional utilization credit, fairly than as a part of the month-to-month Claude subscription tiers (which some have likened to an "all-you-can eat" buffet, making the economics difficult for Anthropic when energy customers and third-party harnesses like OpenClaw eat greater than the $20 or $200 month-to-month consumer spend on the plans in tokens).
OpenClaw’s creator, Peter Steinberger, was notably employed by OpenAI in February 2026 to guide their private agent technique, and has, since becoming a member of, actively spoken out in opposition to Anthropic's limitations — advising that OpenAI's Codex and fashions typically don't have the identical restrictions as Anthropic is now imposing.
By hiring Steinberger and subsequently launching a Professional tier that gives the high-volume capability Anthropic lately restricted, OpenAI is successfully courting the displaced OpenClaw group to reclaim the skilled developer market.
