TypeSafe AI's security overseer moves to GPT Luna foundations, Jev up to 18× faster on p95

TypeSafe AI's security overseer currently runs on the GPT Luna model and automatically analyzes prompts by default; the Jev component is up to 18 times faster and more accurate on p95 latency. The update will be available soon on the Vercel AI Gateway and will likely become the new default, which could increase the system's security and performance.