Anthropic says defenses cut Claude reasoning token theft by up to 99%

What happened

Anthropic says new defenses reduced reasoning tokens extracted from Claude by 90 to 99 percent amid large-scale distillation attacks using hundreds of thousands of fake accounts. It alleges that Zhipu extracted cyber capabilities for GLM-5.3 and that Moonshot secretly routed queries through Claude. Its protections restrict access to internal reasoning, detect extraction attempts with a classifier, and automatically ban flagged accounts; tens of thousands have been banned, with some legitimate users also affected.

The story so far

Anthropic says new defenses reduced the volume of internal reasoning tokens extracted from Claude by 90–99% and led to tens of thousands of account bans, with some legitimate users also affected. The protections restrict access to Claude’s internal reasoning, use a classifier to detect extraction attempts and automatically ban flagged accounts. Anthropic describes the extraction as part of large-scale distillation attacks using hundreds of thousands of fake accounts. It alleges that Zhipu extracted cybersecurity capabilities for GLM-5.3 and that Moonshot secretly routed queries through Claude.

Keep reading

Trending now

1. 1Meta Releases Muse Spark Open With Mac and Cloud Support, Topping US Open-Weight Index
2. 2Google Launches Gemini 4 Argon Globally, Expands Paid Access as Benchmark Results Vary
3. 3OpenAI Updates GPT-6 Astra Pro in ChatGPT With Parallel Agents and Faster Responses
4. 4Meta’s New Business Terms Raise Dispute Over Website Crawling and AI Training Rights