Ever since the new $100 Pro plan, they now claim there's a "dynamic usage limits" that can become restricted at anytime, and not reset for indefinitely as long as they deem it "appropriate"

OpenAI's new $100/month GPT-5.4 Pro subscription is embroiled in controversy as users report sudden, opaque restrictions under a newly cited 'dynamic usage limits' policy. The core complaint is a lack of transparency: customers, including one who shared their experience on Reddit, were shown messages promising a reset date (e.g., April 18th), only for that message to disappear without the limits being restored. When contacting support, users are informed that limits can be restricted indefinitely based on what OpenAI deems 'appropriate' usage, a policy not explicitly detailed on the service's public website or pricing pages.

This incident highlights a significant trust and communication gap between OpenAI and its professional user base. The move to a 'Pro' tier suggested a shift towards more predictable, high-volume access for developers and power users. However, the implementation of vague, retroactively applied limits undermines that premise. For professionals building applications or workflows on GPT-5.4, unpredictable throttling can disrupt operations and planning, making the service unreliable for business-critical tasks. The situation has sparked accusations of 'gaslighting,' as the changing reset messages and undisclosed policy feel like a bait-and-switch to the paying community.