As the debate over AI safety intensifies, third-party evaluation organizations such as Model Evaluation and Threat Research (METR) and Apollo Research are gaining increasing attention. Major AI companies like Anthropic and OpenAI have sought support from these organizations, and President Donald Trump has also encouraged AI companies to collaborate with independent external auditors. However, issues regarding the funding, access, and reporting structure of these evaluation organizations remain unresolved. Last week, OpenAI dismissed three employees, two of whom believe the reason was related to their communication with third-party evaluation organizations, raising concerns about the independence of these organizations. OpenAI denied this claim, stating that it is actively finalizing contracts with third-party safety evaluation organizations. Evaluation organizations like Vals AI and METR have secured tens of millions of dollars in funding this year.