OpenAI Restores Five-Hour Limit After Temporary Removal as GPT-5.6 Sol's 18% Efficiency Gain Sparks Power User Backlash

OpenAI officially announced around July 12, 2026, the temporary removal of the five-hour usage limit for ChatGPT Work and Codex on Plus, Business, and Pro plans, simultaneously resetting all user quotas and releasing GPT-5.6 Sol improvements. Official data shows an 18% efficiency gain and optimized tool calls. Several days later, after active users exceeded 8 million, the five-hour limit was reinstated.

Event Cause and Operational Details

This adjustment was announced by Thomas "Tibo" Sottiaux, who oversees Codex. The announcement explicitly listed three simultaneous actions: removal of the five-hour limit, quota reset, and model efficiency improvements. The statement emphasized the word "temporary," indicating it was not a permanent policy. Sources indicate that ChatGPT Work and Codex share the same five-hour window between local messages and cloud chats. Under the Plus plan, GPT-5.6 Sol heavy tasks are limited to 15 to 90 messages, and the Business plan per-seat quota is identical to Plus.

The quota reset gave users in the middle of their cycle additional allowances, temporarily alleviating usage anxiety for power users. GPT-5.6 Sol's 18% efficiency improvement means fewer quotas are consumed for the same task, and the official confirmation of tool call optimizations directly impacts the shared window mechanism.

Business Logic and Technical Trade-offs

OpenAI's move stemmed from surging demand. After the launch of GPT-5.6 Sol, usage rose rapidly. Temporarily removing the limit allowed testing of infrastructure capacity while reducing per-task quota consumption through efficiency improvements. The reinstatement of the limit, driven by user numbers exceeding 8 million, shows the company prioritized service stability over unlimited supply.

The five-hour window design combines local CLI, IDE, and cloud chat usage to prevent users from bypassing the limit by switching between endpoints. The Business plan's quota being identical to Plus means enterprise users face the same quota pressure as individual users.

Actual Impact on Stakeholders

For developers, the temporary removal provided a higher-density testing window, and GPT-5.6 Sol's 18% efficiency gain reduces quota consumption for consecutive tasks. However, after the limit was restored, workflows relying on high-frequency calls must be replanned. Enterprise users face a reassessment of subscription value: at the same price, the Business plan does not offer higher per-seat quotas, and power users complain that the returned limit directly affects productivity planning.

For OpenAI itself, the move temporarily relieved infrastructure pressure, but long-term balancing between efficiency improvements and capacity expansion remains. Competitors like Anthropic offer longer usage times for paid users during the same period, creating a direct contrast.

Strategic Judgment and Observational Signals

Based on available facts, the most likely development is that OpenAI will continue to alleviate quota pressure through model efficiency iterations rather than immediately expanding capacity. Developers and enterprises should watch whether the next model update comes with new quota adjustments and whether the Business plan gets a separate per-seat quota increase.

In terms of selection advice, teams relying on high-frequency hybrid use of local and cloud should first test GPT-5.6 Sol's quota consumption in actual tasks before deciding whether to maintain current subscriptions or supplement with other providers. Enterprise procurement must clearly distinguish the actual per-seat quota differences between Plus and Business plans.