In the past year, new models from industry leaders have continued to boost AI’s capabilities. According to various capabilities tests, Anthropic’s Mythos model has leapfrogged other AI models – including in the ability to thwart or support cyberattacks.