ChatGPT - sudden rate limits for Plus users in ChatGPT web?
Paid subscribers report hitting unexpected caps, with all models blocked despite being under the 3,000-message weekly limit.
OpenAI's ChatGPT Plus service, a $20/month subscription, is facing user backlash as subscribers encounter sudden and unexplained rate limit errors on the web platform. According to the official help page, Plus users are allotted 3,000 messages per week specifically for the GPT-5.4 Thinking model. However, numerous users on forums like Reddit report hitting a hard cap that blocks access to all models, including the faster 'Instant' option, despite being confident they have not exceeded the published weekly message limit. The error message simply states the user has reached a limit, offering no breakdown of usage or clear path to resolution.
This incident has sparked confusion and frustration among the professional user base, who pay for predictable, uninterrupted access. The discrepancy suggests either a significant technical bug in OpenAI's rate-limiting system or an unannounced, retroactive change to the service's terms. The lack of immediate communication or transparency from OpenAI regarding the cause has left users unable to complete work and questioning the value proposition of the Plus tier. For now, the only workaround appears to be waiting for the perceived limit to reset, leaving many to speculate if this is a temporary glitch or a new, stricter enforcement of usage policies.
- ChatGPT Plus users report being blocked from all AI models, including 'Instant', due to sudden rate limits.
- The official policy states a 3,000-message weekly limit for GPT-5.4, but users are hitting caps much earlier.
- The lack of transparency from OpenAI has caused workflow disruptions and confusion among paying subscribers.
Why It Matters
Paid professionals rely on consistent API access; unexpected downtime disrupts critical workflows and erodes trust in service reliability.