Do New Frontier LLMs Really Resolve Ambiguity Better?
Frontier labs increasingly claim their models understand intent with less prompting. Here is what the evidence supports, what training changed, and how to test ask-versus-guess behavior in your own agents.
21 min read