White House AI Framework Expands to Open Source Models Matching GPT-5.6

US administration tightens AI oversight, extending federal security testing to open source models once capabilities match Anthropic Mythos and OpenAI GPT-5.6 benchmarks.

Woofun AI reports that the Trump administration intends to revise artificial intelligence guidelines to enforce stricter oversight on AI models. The White House has instituted a framework mandating that US laboratories undergo federal security testing prior to public release. In the coming months, this regulatory scope is expected to encompass open source models, requiring testing once their capabilities align with Anthropic's Mythos and OpenAI's GPT-5.6.