tech
White House quiet on OpenAI's Mythos-like model
The seemingly straightforward model release raises questions about what actually triggered the Trump administration's concerns about Anthropic.
TL;DR
- OpenAI released an update to its cybersecurity model, GPT-5.5-Cyber.
- GPT-5.5-Cyber scored 85.6% on the CyberGym benchmark, outperforming Mythos 5's score of 83.8%.
- OpenAI's release did not face the political pushback or export control issues that Anthropic encountered with its models.
- The article questions the reasons behind the Trump administration's concerns regarding Anthropic's models.
- OpenAI has expanded partnerships with organizations in Australia, Canada, France, Germany, Japan, Poland, South Korea, and the EU.