Around 85% success rate of keystroke recovery with our script :)
Britain's cyber agency is warning that AI-fuelled bug hunting is about to flush out years of buried flaws, leaving defenders scrambling to keep up.…
TL;DR: If a large model finds a 0-day with 90% probability, and a small model with 50% probability, but the small model costs 10x less, it is better to use the small model.
We compared the cost and recall of various models in finding real, recent zero-days and found that for most applications, smaller models run repeatedly can significantly outperform larger frontier models on cost-to-recall.
Disclaimer: I'm involved with Hacktron, the company that produced this research. This is a factual presentation of our benchmarks, which we hope the community can use to make informed decisions about models like Mythos.
Login