Anthropic Reduces Claude's Reasoning Token Theft by Up to 99%

Anthropic announced it has reduced the theft of internal reasoning tokens in Claude by 90 to 99%. This measure comes as part of the company's efforts to combat large-scale distillation attacks attributed to Chinese firms Zhipu and Moonshot. The actions taken by Anthropic include banning tens of thousands of accounts linked to these activities. This move reflects the company's commitment to safeguarding its technologies from illicit exploitation, enhancing its competitive edge in the market.