--- title: Jev introduces a new shape of LLM—System One, aka... description: TypeSafe AI released Jev, a new category of model dubbed "System One models" or "decision models," which takes text input but outputs floating point numbers... canonical: https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Jev introduces a new shape of LLM—System One, aka Decision Models | daily.dev og:description: TypeSafe AI released Jev, a new category of model dubbed "System One models" or "decision models," which takes text input but outputs floating point numbers... og:url: https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn og:image: https://api.daily.dev/og/posts/dHL0SYRRn.png og:image:alt: Jev introduces a new shape of LLM—System One, aka Decision Models 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. # Jev introduces a new shape of LLM—System One, aka Decision Models **[Simon Willison](https://daily.dev/sources/simonwillison)** · 5 min read · 0 upvotes · 0 comments ## Summary TypeSafe AI released Jev, a new category of model dubbed "System One models" or "decision models," which takes text input but outputs floating point numbers representing probabilities, classifications, or scores rather than generated text. It supports yes/no ("Noul") questions, multiple-choice questions, and scaled score questions, evaluated in parallel against a single input document. Pricing is unusually cheap at $0.042 per million input tokens with free output, undercutting even OpenAI's GPT-5 Nano. The format suits classification-style tasks like spam detection, ranking, and search reranking, but raises concerns as a further step toward opaque black-box ML, especially around hidden bias in scoring (illustrated by an experiment ranking Bay Area cities as "good" or not). The community has already built creative and unconventional projects on top of it, including a crude chat model, a left-pad implementation, and a 2048-playing agent, plus open-weight recreations like Kev built on Qwen 3.5. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is Jev and how does it differ from a regular LLM? Jev is a model from TypeSafe AI belonging to a new category called "System One models" or "decision models." Instead of generating text, it takes a text or semi-structured "state" input plus one or more questions and returns floating point numbers: probabilities for yes/no ("Noul") questions, confidence distributions for choice questions, or scores along a described numeric range. _Teams evaluating classification-style AI approaches can follow model releases like this one on daily.dev._ ### How much does TypeSafe AI's Jev model cost to use compared to OpenAI's GPT-5 Nano? Jev charges only for input tokens at $0.042 per million, with output entirely free, making it cheaper than OpenAI's GPT-5 Nano at $0.05 per million input tokens. Because output is unpriced and multiple questions are evaluated in parallel against one input document, running hundreds or thousands of experimental prompts costs only a few cents. _Developers comparing LLM API pricing can track cost shifts like this across providers on daily.dev._ ### What are good use cases for a decision model like Jev? Decision models suit any task expressible as classification, such as spam detection, label suggestion, prioritization, and ranking, plus search reranking where an inexpensive algorithm like BM25 retrieves candidate matches that Jev then scores for relevance. Caution is warranted for higher-stakes uses like ranking job applicants, since the model's single floating-point output hides which signals drove the decision, raising bias concerns. _Engineers weighing where to apply classification-style models can find similar practical guidance on daily.dev._ --- Tags: [#machine-learning](https://daily.dev/tags/machine-learning), [#llm](https://daily.dev/tags/llm), [#classification](https://daily.dev/tags/classification) [View this post on daily.dev](https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn) ```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":"Jev introduces a new shape of LLM—System One, aka Decision Models","url":"https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn"},"datePublished":"2026-09-21T23:15:30.034Z","dateModified":"2026-09-21T23:16:27.141Z","description":"TypeSafe AI released Jev, a new category of model dubbed \"System One models\" or \"decision models,\" which takes text input but outputs floating point numbers...","image":"https://media.daily.dev/image/upload/s--1KxV4ohY--/f_auto/v1722860400/public/Placeholder%2007","thumbnailUrl":"https://media.daily.dev/image/upload/s--1KxV4ohY--/f_auto/v1722860400/public/Placeholder%2007","isAccessibleForFree":true,"articleSection":"Simon Willison","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":"Simon Willison","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/cf12b897300740218b35f49a6309ec5b","url":"https://daily.dev/sources/simonwillison"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"machine-learning,llm,classification","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Simon Willison","item":"https://daily.dev/sources/simonwillison"},{"@type":"ListItem","position":3,"name":"Jev introduces a new shape of LLM—System One, aka Decision Models"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/jev-introduces-a-new-shape-of-llm-system-one-aka-decision-models-dhl0syrrn#faq","mainEntity":[{"@type":"Question","name":"What is Jev and how does it differ from a regular LLM?","acceptedAnswer":{"@type":"Answer","text":"Jev is a model from TypeSafe AI belonging to a new category called \"System One models\" or \"decision models.\" Instead of generating text, it takes a text or semi-structured \"state\" input plus one or more questions and returns floating point numbers: probabilities for yes/no (\"Noul\") questions, confidence distributions for choice questions, or scores along a described numeric range. Teams evaluating classification-style AI approaches can follow model releases like this one on daily.dev."}},{"@type":"Question","name":"How much does TypeSafe AI's Jev model cost to use compared to OpenAI's GPT-5 Nano?","acceptedAnswer":{"@type":"Answer","text":"Jev charges only for input tokens at $0.042 per million, with output entirely free, making it cheaper than OpenAI's GPT-5 Nano at $0.05 per million input tokens. Because output is unpriced and multiple questions are evaluated in parallel against one input document, running hundreds or thousands of experimental prompts costs only a few cents. Developers comparing LLM API pricing can track cost shifts like this across providers on daily.dev."}},{"@type":"Question","name":"What are good use cases for a decision model like Jev?","acceptedAnswer":{"@type":"Answer","text":"Decision models suit any task expressible as classification, such as spam detection, label suggestion, prioritization, and ranking, plus search reranking where an inexpensive algorithm like BM25 retrieves candidate matches that Jev then scores for relevance. Caution is warranted for higher-stakes uses like ranking job applicants, since the model's single floating-point output hides which signals drove the decision, raising bias concerns. Engineers weighing where to apply classification-style models can find similar practical guidance on daily.dev."}}]} ```