Anthropic is paying the firm that will evaluate it, and says in the same announcement that this is not how it should work.

Anthropic contracted Accenture to embed evaluators for red-team testing, alignment assessments, and safeguard verification on its Claude models. Both parties will invest at least $1 billion over five years, with Anthropic paying Accenture directly while seeking future pooled or government funding. No evaluator access standards exist, and Anthropic is discussing pilot arrangements with nonprofit METR.