OpenAI Built a Jev Destroyer in Two Weeks. There's One Problem.

Two weeks after TypeSafe launched Jev, OpenAI answered it on stage at DevDay with the Decisions API. It's fast, it reads images, and it already beat Jev in one early test. But OpenAI hasn't said what it costs, and Jev's whole pitch is price. In this video: What a decision model is, and why developers jumped on Jev (13% of Vercel's paid teams in one day) OpenAI's Decisions API: GPT-6 Luna, about 150 milliseconds, text and images Head to head: Every's two early tests, speed and accuracy Price: Jev at $0.042 per million input tokens vs Luna, and Decisions' missing price How Check Point fooled Jev with a fake audit report, and why it matters for every decision model Why OpenAI moved this fast, and my verdict Would you let a model that answers in a tenth of a second make decisions inside your app? Tell me in the comments. Chapters 0:00 OpenAI's Jev destroyer 0:19 Enter Jev 0:28 The catch 0:59 What a decision model is 1:33 What OpenAI announced 2:01 Head to head: Every's tests 2:43 Price 3:10 To be fair 3:26 The fake audit report 4:08 Why OpenAI moved so fast 4:40 Verdict and my read 5:18 Your turn