Jev

"Also very important, with Jev I built hot dog not hot dog... You all know how bullish I am on Jev. I think this is like the missing piece in a lot of our AI architecture."

Lacks vision, unlike the Decisions API

"I've been hooked on Jev. I think there's a bunch of use cases for Jev that people are not talking about, underserved. One that I think is really interesting is just UI design. How do you use a decisions API to get you the most highly con..."

Compared against OpenAI's new Decisions API

"Startup Typesafe announces a model called Jev, and Jev they market as so-called system one intelligence... you can ask it to answer in the form of a categorical... And turns out that's a great idea."

Compared to OpenAI's Decisions API and open-source clones

"This new model dropped a couple weeks ago called Jev, which gives you very, very fast structured output responses for questions that can be used as LLM as a judge or classifier... And it's so much cheaper."

Decisions API was faster on some workloads

"More than any other model I've experienced lately, Jev has been the one that has exploded use cases — personal productivity use cases, code use cases, product use cases. It is going to be in the middle of almost everything I build from h..."

Compared to standard LLMs (Claude Opus, GPT-6) which produce text output and are expensive on output tokens; Jev is positioned as a complementary decision layer alongside frontier models like Gemini and Astra.

"I've used it. It's awesome. What I use it for is deciding who in the SaaStr community should meet each other. It's quite good at that. Infinitely so cheap it doesn't even measure it. Ten times faster is the use case, but only for these l..."

Compared to Anthropic/Claude API for cost; positioned as 'system one' thinking vs LLMs' 'system two'; fails for complex reasoning tasks like ranking candidates

"What Jev is made for is to make really fast, high-quality decisions for incredibly cheap. I'm going to monitor everything — the same way that I just set up my emails. I was able to monitor every email and categorize it and create a level..."

Compared to traditional LLMs which cost ~3 cents and take ~10 seconds per decision vs. Jev at ~$0.00004 and ~0.4 seconds; also compared to GPT-6 Astra and Fable 5.1 as different model types

"This is really the key piece out of the whole JEV thing. How can you access data, scrape it, store it, classify it, enrich it, make it much more useful and usable and pipe that into some kind of service — either making a standalone busin..."

Compared to ChatGPT for classification tasks; noted existing LLMs can handle low volume but JEV shines at scale

"Jev is a classifier at its truest being — that's what it is. You define an input, and then the output is a schema. It doesn't generate any text at all. It just gives you the output — the probabilities of those choices. It takes around 20..."

Compared to LLMs like ChatGPT and Claude which are slow, stream text, and cost more; Jev does not reason or generate text, it purely returns probability scores

"I'm actually willing to make that bet, because I'm using Jev a lot and I can already see that if it never changed, maybe I would swap it out, right? But if it gets better than the rest, there's no way I would change... Jev is super fast ..."

Competitive products from Cloudflare, OpenAI, Amazon, and OpenRouter