Checked fact 127913 Oct 2026Safety and security

The post says that, by Anthropic's own estimate of reasoning tokens extracted on its own platform, after it introduced the classifier the number extracted by distillers is "consistently 10 to 100 times lower than it was before."

Toned down to what the source says

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

The exact words it rests on

To measure the effectiveness of our system, we estimate how many tokens of Claude's reasoning distillers have extracted on our own platform. After introducing the classifier, the number of reasoning tokens extracted by distillers is consistently 10 to 100 times lower than it was before

What the source said when we opened it, on 13 Oct 2026.

The source

Strengthening our defenses against distillation attacks
Anthropic · 2026-10-12

Open the source archived copy

Checked

Checked by the notis newsroom on 13 Oct 2026, against the source above.

In the story

Anthropic says new safeguards cut distillers' extraction of reasoning tokens 13 Oct 2026

Cite this fact

Short permanent address

In a sentence

“The post says that, by Anthropic's own estimate of reasoning tokens extracted on its own platform, after it introduced the classifier the number extracted by distillers is "consistently 10 to 100 times lower than it was before.".” ai notis, checked 13 Oct 2026. Source: Strengthening our defenses against distillation attacks (Anthropic). https://ainotis.com/c/1279

As a reference (APA style)

ai notis. (2026, October 13). The post says that, by Anthropic's own estimate of reasoning tokens extracted on its own platform, after it introduced the classifier the number extracted by distillers is "consistently 10 to 100 times lower than it was before." [Checked claim 1279]. ai notis. https://ainotis.com/c/1279

Anyone may quote this address. It does not change; if we correct the story, this page says so.