On Nov 7, multiple users reported on the V2EX community that OpenAI and Anthropic have recently carried out a concentrated wave of account bans targeting users in China. The detection focuses on whether reverse proxies are in use, and the bans have hit subscriptions across low-price regions, Team, Plus and Pro, with even full-price US Pro users caught. As of publication, neither company has issued a public statement on the bans.
Scope of the bans and what is being checked
According to a V2EX post titled "A summary of today's GPT ban wave," the bans "focus on whether reverse proxies are in use, and have nothing to do with full-price versus low-price regions." Low-price regions appear harder hit because they have more users, creating the illusion that "full-price US accounts were spared." The user said that among US full-price Pro accounts at the $200 tier, users who used reverse proxies normally were also banned.
The post also said the bans currently "target only domestic users," with low-price regions and Team the "worst hit," followed by Plus and Pro. Reverse-proxy use triggers linked bans: one person had more than a dozen Pro 200 accounts all banned, "with no exceptions." Other users reported being banned without using reverse proxies, but the authenticity of those claims "remains to be verified."
Claude users hit with portal verification
Another V2EX user posted that their Claude Pro subscription "lasted a month" and their Max 5x subscription "lasted less than a month." After the ban, they were hit with the "hardest 'portal verification,'" and the resellers they approached also "said it couldn't be passed." The user said they were on an old account with no suspicion of distillation, and had opened Max to use Opus 5.5's video capabilities.
I make popular-science videos — does that threaten U.S. national security too? Future AI must be diversified! Diversified!! Whoever dominates, hit them!
User reaction and what to watch
The poster questioned how mature the ban-detection system is, saying "the ban-detection mechanism isn't mature yet, and OpenAI couldn't wait to use it — that's clearly abnormal." The user speculated that either an official announcement explaining the bans will come, or the ban mechanism will be aligned with Anthropic's, noting that "A社's cat joined OpenAI and is the top product manager there, with decision-making power."
The user also warned that "the bans are ongoing — not being killed now doesn't mean you're safe," and said the episode could escalate into a large-scale crackdown on OpenAI use by domestic users, with an official explanation becoming the norm. As of publication, discussion was still ongoing on V2EX and other communities, and the exact scale of affected users and official responses from both companies remain to be confirmed.