OpenAI Removes 5-Hour Limit for Codex and ChatGPT Work

OpenAI has removed the five-hour usage limit that applied to Codex and ChatGPT Work for users on paid plans.

The change took effect in the past 48 hours. It covers Plus, Business, and Pro accounts. The company also reported progress on a new version of its model that reduces token consumption during tasks.

Six million active users now interact with the tools each month. OpenAI stated that usage counters will reset within the next hour for all accounts.

The five-hour restriction had forced many developers to monitor their daily activity closely. Removing it removes one friction point for teams that run long sessions of code generation or document work.

The efficiency update targets GPT 5.6 Sol. The company claims the new version completes the same work with fewer tokens, though exact savings numbers have not been published yet.

This move arrives after several months of capacity complaints on forums and social platforms. Users often hit the limit during extended debugging or research sessions and then faced reduced speed until the next reset window.

OpenAI did not link the removal to any single competitor action. It framed the step as part of ongoing work to increase daily throughput rather than a direct response to market pressure.

The 6 million active user count marks steady growth from earlier public figures. The company did not break down the split between free and paid accounts.

Several analysts noted that lifting hard caps can increase total compute demand. They questioned whether OpenAI has built enough headroom to keep response times stable after the change.

The model efficiency work is separate from the cap removal. OpenAI said GPT 5.6 Sol changes will become visible over coming weeks once internal measurements are complete.

Developers who build agents or automated pipelines stand to gain the most. They can now run longer test loops without watching a usage timer.

Teams inside companies that adopted ChatGPT Work for internal documents face fewer interruptions. A single complex report or code review can now finish without an enforced pause.

It remains unclear how quickly OpenAI will publish the token-reduction numbers for the new model version. The company has delayed similar metrics in past rollouts.

Users should watch average session length and error rates over the next month. Any spike in timeouts would indicate that capacity has not kept pace with higher demand.

The reset timing also matters. If the new window still clusters around a fixed daily boundary, some power users may simply shift their heaviest work to the start of each period.

OpenAI faces parallel pressure from competitors that market unlimited or much higher limits. The current step narrows one visible gap but does not eliminate the comparison.