The quality of 5.6 Sol and Opus 5 / Fable 5 compared to what we have before, especially when it comes to vulnerability analysis is night and day.
Even distilled models like K3 are leagues ahead of what we had for the past 2 years.
If you haven't yet, use a modern Clanker on the highest effort setting to comprehensively audit your code. And give it a harness that let's it experiment and actually test it's theories.
