Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now
Today, many eyes are on OpenAI CEO and co-founder Sam Altman‘s ongoing public feud with Elon Musk on the latter’s social network, X.
But Altman’s recent statements regarding the ongoing rollout of his company’s latest and greatest large language model (LLM), GPT-5, are probably more important to customers and enterprise decision-makers.
After an admittedly “bumpy” debut of GPT-5 last week that saw some users clamoring for restored access to deprecated older LLMs in ChatGPT such as GPT-4o and o3 — OpenAI granted the former — Altman is now pivoting towards ensuring OpenAI’s underlying infrastructure and usage limits are a good fit for the company and its 700 million active weekly ChatGPT users.
The company’s latest updates include a more detailed compute allocation plan and the introduction of additional third-party connectors for ChatGPT Plus and Pro plans.
AI Scaling Hits Its Limits
Power caps, rising token costs, and inference delays are reshaping enterprise AI. Join our exclusive salon to discover how top teams are:
Turning energy into a strategic advantage
Architecting efficient inference for real throughput gains
Unlocking competitive ROI with sustainable AI systems
Secure your spot to stay ahead: https://bit.ly/4mwGngO
Managing GPT-5 demand and usage limits
In a post on X last night, Altman outlined how OpenAI will prioritize computing resources over the next several months.
Here is how we are prioritizing compute over the next couple of months in light of the increased demand from GPT-5:1. We will first make sure that current paying ChatGPT users get more total usage than they did before GPT-5.2. We will then prioritize API demand up to the…— Sam Altman (@sama) August 12, 2025
He said the company’s first priority is ensuring that current paying ChatGPT users receive more total usage than they had before GPT-5’s release, though he did not provide specific figures for the increase.
However, Altman previously posted on X that OpenAI was “trying” a 3,000 messages-per-week usage limit when using the GPT-5 “thinking” mode, more reasoning power and time spent reasoning on harder problems, for ChatGPT Plus subscribers (the $20 per month plan).
Interestingly, one report from an AI app creator on X said that OpenAI told him the usage limits for GPT-5 plus thinking on the ChatGPT Team plan ($30 per user per month) is much lower than that of ChatGPT Plus users, only 200 “Thinking” messages per week when selected manually by the user.
OpenAI just replied to me with an email about the GPT-5 usage limits under the Team plan：· ChatGPT Team can manually select GPT-5-Thinking· Manual usage cap: 200 messages/week· On reaching cap: popup alert, GPT-5-Thinking hidden from menu— Vic Zhang (@RealVicHere) August 12, 2025
OpenAI’s availability of GPT-5 through its application programming interface (API) for third-party developers is also being tweaked.
Altman also stated in his X post that OpenAI would “prioritize API demand up to the currently allocated capacity and commitments we’ve made to customers.”
In other words, existing API users and those already in contract will get the first dibs on GPT-5 access through OpenAI’s API, others may have to wait longer.
Altman also clarified “we can support about an additional ~30% new API growth from where we are today with this capacity,” meaning they can take on more API users, but not too many.
While OpenAI hasn’t definitiv …