Anthropic's two-stage anti-distillation classifier runs on Claude Sonnet 5.5, Claude Opus 5 and 5.5, and Claude Fable 5 and 5.1, with a false positive rate below 0.1%.

● OPEN · settles 2026-10-20

Background

Anthropic says the quiet copying of frontier models has become an industrial operation, run through proxy networks holding hundreds of thousands of fraudulent accounts, and that some of what gets extracted ends up in foreign military and surveillance work. Its answer is to treat extraction the way it treats requests for help with bioweapons. A two-stage classifier now screens Claude's most capable models for attempts to pull out the reasoning behind an answer, and a change called preserved thinking stops new API accounts from editing earlier context to pry transcripts loose. The company says reasoning tokens extracted on its own platform have fallen 10 to 100 times since the classifier arrived, that it has banned tens of thousands of accounts, sometimes within two hours of a campaign starting, and that it now charges for blocked requests in the categories where its false positive rate sits below 0.1%. It is also naming names: Zhipu, it says, distilled cyber capabilities while building GLM-5.3, and requests sent to Moonshot's Kimi models were secretly forwarded to Claude, one of them a surveillance query it links to China's People's Liberation Army.

This market exists because launch announcements are easy to make and hard to grade. Press releases describe what a company intends to ship; they rarely come back later to say whether it shipped. AI Rank Bank pulls one checkable sentence out of the coverage — the sentence above — gives it a date, and lets readers price it in public. The launch itself stays with its publisher; what lives here is the claim, the crowd's call, and eventually the receipt.

How this claim is judged

Judging criteria, fixed when the market opened: Delivered if, at the check time, Anthropic's October 12 post or its platform documentation still state that the two-stage anti-distillation classifier runs on Claude Sonnet 5.5, Claude Opus 5 and 5.5, and Claude Fable 5 and 5.1, with a false positive rate below 0.1%. Anthropic withdrawing the classifier, narrowing that model coverage, or restating the false positive rate at 0.1% or higher does not count.

The source of record is the publisher's own page linked above. A claim about a company's product is judged against that company's public output — its product pages, documentation, changelogs and official posts — not against rumors or third-party summaries. Only yes-or-no outcomes count: partial credit is not a verdict, and a claim that needs a paragraph of excuses has already answered itself.

How readers called it

1 reader called this claim. 100% said it would deliver.

1 Will deliver · 0 Won't ship

Points reward being right against the crowd. A correct call pays 100 × (1 − the share of readers who agreed with you when you placed it), with a floor of 10: side with a 90% consensus and a win pays 10, stand alone and a win pays up to 100. Wrong calls pay nothing. The price is locked in when the call is placed, so early readers are paid for judging before the crowd forms, not for copying it.

How readers split

Share of readers saying “will deliver”: opened at 100%, now at 100%. 3 snapshots on record since 2026-10-13 — a snapshot lands whenever a call is placed, and at least hourly while the market is open.

Verdict

No verdict yet. This claim is checked on 2026-10-20 08:00 Eastern Time (12:00 UTC) against the source of record below. When the check happens, one of three stamps lands on this page permanently: delivered if the claim came true as stated, busted if it did not, or unproven if the date passes without public evidence either way. Unproven claims pay no points and do not count against any reader's hit rate.

About this market

AI Rank Bank is an AI news front page with a memory. Every launch on the homepage carries one claim like this one, with a check date. Readers call claims with points, build a public record, and climb the leaderboard. Companies build a record too: every settled claim lands on the company's bank page, where its delivery rate is simple arithmetic that nobody gets to edit afterwards. There is no money in the market — points buy bragging rights only, which keeps the incentives pointed at being right.