Several frontier AI labs signed the September 2026 White House Accord on Super Intelligence, a voluntary pledge committing signatories including Anthropic, OpenAI, Google, Meta, and xAI to partner with independent external evaluators for model assessments alongside internal controls and board oversight. Anthropic has moved fastest, embedding METR researchers with employee-like access and publication rights in mid-October while detailing the process on October 31; OpenAI completed a METR pilot review identifying monitoring flaws and is finalizing additional contracts. These steps reflect competitive positioning around AI safety credibility amid voluntary frameworks rather than mandates, with CAISI voluntary testing already covering all major U.S. labs. Trader sentiment hinges on whether additional labs announce concrete third-party arrangements before resolution today, as the accord lacks enforcement and timelines remain flexible.