• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Jev arrives on OpenRouter in beta

    OpenRouter says Typesafe's Jev returns structured decisions with probabilities rather than generated text. A browser-agent creator reports using it with Browser Use to find flights in 7 seconds for $0.0039.

    Sebastian RaschkaSR
    Aaron LevieAL
    Yuntian DengYD
    76 Sources, ,

    TLDR

    OpenRouter announced Jev's beta availability on September 18, 2026. It says Typesafe's model takes an app's state plus a typed question and returns a typed decision with a probability attached, rather than generating text. Reported uses include checking citations, ranking news feeds and investigating agent failures. A browser-agent creator says finding flights with Browser Use and Jev took 7 seconds and cost $0.0039. Beyond those tests, one user proposes leaving generation to a language model while Jev handles judgment calls: choosing models or tools, deciding whether to retry and flagging tasks for human review.

    Combined views

    2.9M

    76 Sources, first seen 21d ago

    Combined views

    2.9M

    76 Sources, first seen 21d ago

    27.6K likes
    21d ago
    first seen 21d ago
    27.6K likes
    1.4K comments
    18.6K saves
    2.3K reposts
    1.4K comments
    18.6K saves
    2.3K reposts

    Sentiment

    Positive71%29%Negative

    Summary

    Positive accounts welcomed Jev for enabling clear, useful harnesses and better generalization in classification tasks, while negative replies dismissed it as lacking real ML progress or warned against building such tools.

    Based on 287 sentiment-bearing replies from 262 accounts across 5 conversations.

    Sentiment

    Positive71%29%Negative

    Summary

    Positive accounts welcomed Jev for enabling clear, useful harnesses and better generalization in classification tasks, while negative replies dismissed it as lacking real ML progress or warned against building such tools.

    Based on 287 sentiment-bearing replies from 262 accounts across 5 conversations.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    76 Sources

    Antaripa Saha@doesdatmaksensegot access to Jev!! heard a lot of good things about it, especially as a classifier. really curious to try it for evals. lots of llm-as-a-judge for binary decisions are going to get replaces. give Jev some state + a bunch of questions and get back probabilities you can actually threshold on. also feels pretty useful inside agent harnesses as a control plane for making all those fuzzy decisions like which model or tool to route to, whether to continue or retry, whether an output looks good enough, whether something needs human review, etc. basically letting the llm handle generation, while Jev handles a lot of the judgment calls around it.21d
    Shreya Shankar@sh_reyaRT @doesdatmaksense: got access to Jev!! heard a lot of good things about it, especially as a classifier. really curious to try it for eva…20d
    Hamel Husain@HamelHusainRT @isaac_flath: I've been using Jev. Here's the six things i've tried and am confident I'll still use Jev for 60 days from now. There's…20d
    Isaac Flath@isaac_flathI've been using Jev by @typesafeai Here's the six things i've tried and am confident I'll still use Jev for 60 days from now. There's many more experiments, ideas, and things I think I will use it for. It's a big deal (more on why in next post). But I am only sharing things that I am 99% sure will lead to stuff I will still be using Jev for in 60 days. That means I started with small, boring, but useful, stuff. - Fact-checking my scripts - Ranking my news feed - Finding the right text in PDFs - Checking citations - Grouping my review notes - Figuring out why agents fail (eval over traces) https://isaacflath.com/writing/six-things-i-tried-with-jev20d
    Pietro Schirano@skiranoJev is actually really really good wow20d
    Bryan Bischof fka Dr. Donut@BEBischofA lot of people are excited about jev the model. Some people are excited about the launch video. But watching Diogo handle the reactions is arguably the most impressive thing about the last two days. No bad vibes, no defensiveness, just sticking to the facts.20d
    elvis@omarsar0Good take! After testing it, Jev feels like an important primitive for building reliable AI systems. I think a few more primitives are waiting to be discovered that could make LLM-based agents even better and faster.20d
    david fant@da_fantjev will make agents 10x faster and cheaper, here's how: 1/ model routing: pick the right model for each task, without training a custom router 2/ computer use: faster, cheaper and more reliable for action-heavy tasks 3/ auto review: ask jev whether an action is safe, instead of using a slow and expensive LLM 4/ less obvious: subagent orchestration long-running agents (cursor projects, grokbot, energy) parallelize work with subagents. but every user message, email, or subagent reply can wake the expensive orchestrator. example: it costs $1 to wake up gpt 6 astra w 100k input tokens jev can decide what each event needs: - route directly to a subagent - queue for later - wake the orchestrator20d
    María Benavente@merybenaventeyou know what else Jev helps classify? people who actually understand ML, and people who don't.20d
    gabriel@gabriel1RT @da_fant: jev will make agents 10x faster and cheaper, here's how: 1/ model routing: pick the right model for each task, without traini…20d

    76 Sources

    Antaripa Saha@doesdatmaksensegot access to Jev!! heard a lot of good things about it, especially as a classifier. really curious to try it for evals. lots of llm-as-a-judge for binary decisions are going to get replaces. give Jev some state + a bunch of questions and get back probabilities you can actually threshold on. also feels pretty useful inside agent harnesses as a control plane for making all those fuzzy decisions like which model or tool to route to, whether to continue or retry, whether an output looks good enough, whether something needs human review, etc. basically letting the llm handle generation, while Jev handles a lot of the judgment calls around it.21d
    Shreya Shankar@sh_reyaRT @doesdatmaksense: got access to Jev!! heard a lot of good things about it, especially as a classifier. really curious to try it for eva…20d
    Hamel Husain@HamelHusainRT @isaac_flath: I've been using Jev. Here's the six things i've tried and am confident I'll still use Jev for 60 days from now. There's…20d
    Isaac Flath@isaac_flathI've been using Jev by @typesafeai Here's the six things i've tried and am confident I'll still use Jev for 60 days from now. There's many more experiments, ideas, and things I think I will use it for. It's a big deal (more on why in next post). But I am only sharing things that I am 99% sure will lead to stuff I will still be using Jev for in 60 days. That means I started with small, boring, but useful, stuff. - Fact-checking my scripts - Ranking my news feed - Finding the right text in PDFs - Checking citations - Grouping my review notes - Figuring out why agents fail (eval over traces) https://isaacflath.com/writing/six-things-i-tried-with-jev20d
    Pietro Schirano@skiranoJev is actually really really good wow20d
    Bryan Bischof fka Dr. Donut@BEBischofA lot of people are excited about jev the model. Some people are excited about the launch video. But watching Diogo handle the reactions is arguably the most impressive thing about the last two days. No bad vibes, no defensiveness, just sticking to the facts.20d
    elvis@omarsar0Good take! After testing it, Jev feels like an important primitive for building reliable AI systems. I think a few more primitives are waiting to be discovered that could make LLM-based agents even better and faster.20d
    david fant@da_fantjev will make agents 10x faster and cheaper, here's how: 1/ model routing: pick the right model for each task, without training a custom router 2/ computer use: faster, cheaper and more reliable for action-heavy tasks 3/ auto review: ask jev whether an action is safe, instead of using a slow and expensive LLM 4/ less obvious: subagent orchestration long-running agents (cursor projects, grokbot, energy) parallelize work with subagents. but every user message, email, or subagent reply can wake the expensive orchestrator. example: it costs $1 to wake up gpt 6 astra w 100k input tokens jev can decide what each event needs: - route directly to a subagent - queue for later - wake the orchestrator20d
    María Benavente@merybenaventeyou know what else Jev helps classify? people who actually understand ML, and people who don't.20d
    gabriel@gabriel1RT @da_fant: jev will make agents 10x faster and cheaper, here's how: 1/ model routing: pick the right model for each task, without traini…20d