--- title: The fastest-adopted model on Vercel this year cannot... description: TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, launched a model called Jev that never generates text — it only returns structured decisions (a... canonical: https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The fastest-adopted model on Vercel this year cannot write a word | daily.dev og:description: TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, launched a model called Jev that never generates text — it only returns structured decisions (a... og:url: https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s og:image: https://api.daily.dev/og/posts/hyJY7pn9S.png og:image:alt: The fastest-adopted model on Vercel this year cannot write a word og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # The fastest-adopted model on Vercel this year cannot write a word **[The Next Web](https://daily.dev/sources/tnw)** · 5 min read · 0 upvotes · 0 comments ## Summary TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, launched a model called Jev that never generates text — it only returns structured decisions (a choice, a score, or a probability) with calibrated confidence. Released into early access on 15 September alongside a $40m DCVC-led seed round, Jev reached nearly 13% of Vercel's paid teams within 24 hours on Vercel's AI Gateway, the fastest adoption of any model there, beating GPT-5.6 and Fable 5.1. Vercel reported Jev made its safety-classifier step 5–18x faster than using OpenAI's Luna 5.6. TypeSafe charges $0.042 per million input tokens and claims 70-500ms response times versus 3-329 seconds for frontier models on comparable queries, though it has disclosed several limitations in its own benchmarking methodology and has not revealed Jev's underlying architecture. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is TypeSafe's Jev model and how is it different from a normal LLM? Jev is a decision-only AI model from TypeSafe AI that never generates text; it takes program state plus typed questions and returns a choice (up to 255 options), an ordered score, or a yes-no probability, each with a calibrated confidence number. It cannot chat, draft, explain, or write code, but responds in 70-500 milliseconds versus 3-329 seconds for frontier language models on comparable queries, and charges $0.042 per million input tokens with no output metering. _Anyone weighing decision-only models against LLMs for agent pipelines can track this shift on daily.dev._ ### How fast was Jev adopted on Vercel's AI Gateway compared to other models? Jev reached nearly 13% of Vercel's paid teams within 24 hours of launching on Vercel's AI Gateway, twice the share reached by the GPT-5.6 family and more than six times that of Fable 5.1 in the same window. Every other recent launch had stayed below 7% after a full day, making it the fastest-spreading model Vercel had seen on the gateway. _Developers deciding which new model to adopt into their stack can follow adoption trends like this on daily.dev._ ### What tasks did Vercel replace with the Jev decision model instead of a language model? Vercel replaced a safety classifier running on OpenAI's Luna 5.6, which checked whether agent commands were safe to execute, with TypeSafe's Jev model. The swap made that classification step five to eighteen times faster, since the task was a routing or scoring decision that never actually needed free-text generation. _Teams optimizing agent guardrails and routing costs can keep up with cases like this on daily.dev._ --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#vercel](https://daily.dev/tags/vercel) [View this post on daily.dev](https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"The fastest-adopted model on Vercel this year cannot write a word","url":"https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s"},"datePublished":"2026-09-22T16:00:12.488Z","dateModified":"2026-09-22T16:01:16.841Z","description":"TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, launched a model called Jev that never generates text — it only returns structured decisions (a...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5d8e274193e036e168e736ac3afebc8f?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5d8e274193e036e168e736ac3afebc8f?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The Next Web","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The Next Web","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/tnw","url":"https://daily.dev/sources/tnw"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-agents,vercel","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The Next Web","item":"https://daily.dev/sources/tnw"},{"@type":"ListItem","position":3,"name":"The fastest-adopted model on Vercel this year cannot write a word"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-fastest-adopted-model-on-vercel-this-year-cannot-write-a-word-hyjy7pn9s#faq","mainEntity":[{"@type":"Question","name":"What is TypeSafe's Jev model and how is it different from a normal LLM?","acceptedAnswer":{"@type":"Answer","text":"Jev is a decision-only AI model from TypeSafe AI that never generates text; it takes program state plus typed questions and returns a choice (up to 255 options), an ordered score, or a yes-no probability, each with a calibrated confidence number. It cannot chat, draft, explain, or write code, but responds in 70-500 milliseconds versus 3-329 seconds for frontier language models on comparable queries, and charges $0.042 per million input tokens with no output metering. Anyone weighing decision-only models against LLMs for agent pipelines can track this shift on daily.dev."}},{"@type":"Question","name":"How fast was Jev adopted on Vercel's AI Gateway compared to other models?","acceptedAnswer":{"@type":"Answer","text":"Jev reached nearly 13% of Vercel's paid teams within 24 hours of launching on Vercel's AI Gateway, twice the share reached by the GPT-5.6 family and more than six times that of Fable 5.1 in the same window. Every other recent launch had stayed below 7% after a full day, making it the fastest-spreading model Vercel had seen on the gateway. Developers deciding which new model to adopt into their stack can follow adoption trends like this on daily.dev."}},{"@type":"Question","name":"What tasks did Vercel replace with the Jev decision model instead of a language model?","acceptedAnswer":{"@type":"Answer","text":"Vercel replaced a safety classifier running on OpenAI's Luna 5.6, which checked whether agent commands were safe to execute, with TypeSafe's Jev model. The swap made that classification step five to eighteen times faster, since the task was a routing or scoring decision that never actually needed free-text generation. Teams optimizing agent guardrails and routing costs can keep up with cases like this on daily.dev."}}]} ```