Anthropic says it cut Claude reasoning token theft by up to 99% in fight against industrial-scale distillation
90–99% drop in reasoning token theft. Anthropic banned tens of thousands of accounts linked to Zhipu and Moonshot's industrial-scale distillation. Here's what changed.
Why it matters
Anthropic disclosed a major defensive win against coordinated model distillation attacks targeting Claude's reasoning tokens. The scale and attribution matter: this confirms a known threat vector is operationalized at industrial volume by named Chinese labs, and Anthropic's countermeasures—account bans and technical hardening—are measurable. Practitioners running Claude in enterprise environments need to understand the attack surface; the distillation risk affects cost and model-moat strategy.
The key facts
7 to knowAnthropic reports 90–99% reduction in reasoning token extraction
Attacks attributed to Zhipu AI and Moonshot AI (named Chinese companies)
Tens of thousands of accounts banned by Anthropic
Industrial-scale distillation campaign targeting Claude's internal reasoning tokens
Defensive measures include account-level controls and technical hardening (mechanism not detailed)
No independent verification of attack volume or mitigation effectiveness provided
No disclosure of financial impact, token volumes stolen pre-mitigation, or timeline of attack detection
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Anthropic says it has cut the extraction of Claude's internal reasoning tokens by 90 to 99 percent. The company links the large-scale distillation attacks to Chinese companies Zhipu and Moonshot, and has already banned tens of thousands of accounts.