In June 2026, a large number of posts about ChatGPT and Codex accounts being banned simultaneously appeared on Chinese developer communities such as V2EX. Some accounts were normally paid, normally used, and had no trace of scripts, but were also caught in a one-size-fits-all sweep.

Most account bans have traces, but "false bans with a one-size-fits-all approach" do exist. First, figure out which category you fall into, then decide whether to spend time appealing or switch channels.
Many people think that ChatGPT is a web product and Codex is a developer tool, and that problems with one should not affect the other. In fact, they share the same OpenAI account system — after an account is disabled, you cannot log in to ChatGPT web or Codex CLI, and the API keys under the account also become invalid.
6 Common Reasons for Account Bans
Common Reasons (triggered by both ChatGPT and Codex)
Abnormal Network Environment
The most common one. OpenAI's risk control scores IP reputation, and the following network patterns will continuously deduct points:
Shared proxy egress (hundreds or thousands of people egressing from the same IP at the same time, one user's violation penalizes the entire pool) Frequent cross-country hops (the same account showing three egress IPs from Japan, Singapore, and the US within 30 minutes) Known data center IP ranges (AWS / GCP / DigitalOcean and other public cloud egresses marked as 'non-residential') Proxy services already flagged by the risk control database (SMS activation platforms, bulk