Claude Account Bans, Suspensions and Appeals

Anthropic accuses GLM5.3 and Kimi of distilling Claude's thinking data and says it has banned tens of thousands of related accounts, some within two hours.

Anthropic says proxy networks ran hundreds of thousands of fake accounts, one lab sent 12,000 requests to extract reasoning, tens of thousands of accounts were banned often within two hours, and under 0.1% of legitimate requests were wrongly flagged.

The author says an old Claude account was banned after repeatedly maxing the 5-hour usage limit, with an appeal unanswered, and advises leaving headroom and building accounts in tiers.

A reply claims that amid Claude account bans, users who can still access it are leaving, and those staying either genuinely like it or can't be bothered to switch.

A user whose Claude account was banned jokes about asking whether a seller could pull strings to get the account unbanned after a purchase.

After a Claude ban, the author weighs using a cheap Xianyu Notion membership to access Opus, concluding it works as a stopgap but is limited for heavy use and risky with shared accounts.

The author shares anti-ban tips: a self-built VPS with a clean IP, clearing local Claude traces, never sharing the account, and keeping weekly usage under 80%, which has kept the account stable for a year.

A reply offering a self-described modern survival guide warns not to run reverse engineering on an official Claude account because it will get you banned.

A post jokingly sums up how power users avoid Claude bans: don't use the desktop app, CLI, mobile app, API, or web client.

A warning post says running the REA tool on Claude and ChatGPT will get accounts banned, and links to its GitHub repository.

A repost jokes that anyone who can keep a Claude account unbanned while known to be in China would be hired on the spot.

A user who says their Claude account was banned like others' posts an image and asks whether people now know how to handle it.