AI brief
Anthropic says it reduced extraction of Claude's internal reasoning tokens by 90 to 99 percent, and banned tens of thousands of accounts. It links the large-scale distillation attacks to Chinese companies Zhipu and Moonshot.
Why it matters: Large-scale distillation of a model's reasoning tokens can let other companies train competing systems on its outputs, which Anthropic is now working to block.
Written by AI from The Decoder's published text.
Background
From our own funding, model and news trackers, not from the source.
Anthropic
- Latest funding
- Growth · $65B · 28 May 2026
- Newest model on the API
- Claude Haiku 5.5 · 7 Oct 2026 · $0.1 in / $0.5 out per million tokens
Moonshot AI
- Newest model on the API
- Kimi K3.1 Pro · 23 Oct 2026 · $2 in / $10 out per million tokens
- Last 30 days
- 1 news story
Source
“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.”
Reporting from a publisher outside the organisation being covered.
Read original story ↗