OpenAI Codex lead points to sub2api as users report shrinking limits

OpenAI’s Codex lead, Thibault Sottiaux, has clarified that the company did not change usage limits without informing developers. 

The company is instead directing the blame to third-party “sub2api” tools after paying users reported that their weekly Codex allowance was draining far faster than the plans they bought.

Did OpenAI reduce its usage limits? 

Thibault Sottiaux, who now runs both ChatGPT and Codex as OpenAI’s head of core products, addressed complaints from developers that their weekly codex allowance was draining quickly in a widely shared X post. 

Sottiaux wrote that adjusting usage caps is not something the company does without talking to the community and being transparent about it. 

When his team looked at the accounts that were burning through allowances, he said that many of them were running “sub2api” setups, which repackage a ChatGPT subscription so it can be called like OpenAI’s metered, pay-as-you-go API. 

However, OpenAI’s own documentation notes, subscription access and API-key access are billed on separate tracks.

On OpenAI’s developer forum, a Codex 20x subscriber posting as anil.c1 wrote on August 14 that he was using up his weekly allowance in an almost five-hour window, despite the fact that he was doing the same amount of work with the basic Codex Desktop and no extra tools. Four days later, he posted again to say he had canceled his plan. 

Another user, “plutavian,” said a brand new 20x account had a weekly limit of about $200, while an older account on the same plan still had over $2,000.

Is the OpenAI usage problem only caused by third-party tools? 

A developer named Ayaan Lashari built a free app called NerfTrack that reads Codex usage data and shows the weekly value in dollar terms, and put it on GitHub. The app helps users since OpenAI only shows a percentage bar. 

An investigation by Kingy AI said the claims by NerfTrack were “supported but unproven.” They pointed out that a smaller allowance and faster spending look exactly the same on a rounded percentage bar, so it is hard to tell which one is happening. 

The report also mentioned an August 20 post by Alex Getman, which claimed a Plus allowance had dropped from about $160 to $80 in API-dollar terms. 

This is the second time this issue has come up in two months, but Sottiaux held a Sunday “warroom” meeting on June 30 to explain that the first incident was due to Codex doing extra work behind the scenes by mistake. 

Automated review tools and helper subagents were sometimes running twice or trying too hard to fix errors, which used up more usage than intended. 

He also said the dashboard showed some activity that was never charged and confirmed that fixes were put in place and everyone’s usage limits were fully reset. 

Then, on July 28, according to Kingy AI, he addressed another round of complaints, saying GPT-5.6 Sol makes more tool calls and runs longer than older models, and that OpenAI had adjusted it to make normal use last about 18% longer. He also denied that they had cut any subscription usage limits.

The real problem is the confusion about how the usage limit works. OpenAI’s help center says Codex, ChatGPT Work, and related tools all share the same allowance. Usage also depends on many factors, like the model being used, where the task runs, how complex it is, as well as the context, reasoning effort, speed, and tools used. 

For example, Kingy AI noted that Fast Mode runs GPT-5.6 about 1.5 times faster but spends credits at 2.5 times the normal rate. 

The smartest crypto minds already read our newsletter. Want in? Join them.