--- title: AWS launches a local answer to TypeSafe’s Jev decision model description: AWS released Strands Decider 2B, a small downloadable decision model built on Qwen3.5-2B that scores developer-supplied answer options instead of generating... canonical: https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: AWS launches a local answer to TypeSafe’s Jev decision model | daily.dev og:description: AWS released Strands Decider 2B, a small downloadable decision model built on Qwen3.5-2B that scores developer-supplied answer options instead of generating... og:url: https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs og:image: https://api.daily.dev/og/posts/UQpPRQNHS.png og:image:alt: AWS launches a local answer to TypeSafe’s Jev decision model 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. # AWS launches a local answer to TypeSafe’s Jev decision model **[The New Stack](https://daily.dev/sources/newstack)** · 4 min read · 2 upvotes · 0 comments ## Summary AWS released Strands Decider 2B, a small downloadable decision model built on Qwen3.5-2B that scores developer-supplied answer options instead of generating free text, intended to give AI agents a fast grounding check before acting. It follows TypeSafe's Jev and competes with OpenAI's hosted Decisions API, Kev, imajev, and Laya. AWS published training data and scripts, reports sub-100ms latency on an RTX 3090 and around 150ms on an M3 MacBook, and says the model ranks first among public models with a full training recipe on JevBench. It integrates with AWS's Strands agent framework via an intervention system that can proceed, deny, request confirmation, or return feedback, with generative calls still routed through Amazon Bedrock. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is AWS Strands Decider 2B and how does it work? Strands Decider 2B is a small decision model AWS released that scores developer-supplied answer options or returns numerical scores instead of generating free-form text. It uses Qwen3.5-2B as its base model, replacing the language-model head with a pointer head of just over a million parameters, trained with a rank-16 LoRA adapter, letting agents check proposed actions before they run. _Teams validating agent actions before execution can follow decision-model releases like this one on daily.dev._ ### How fast is AWS Strands Decider 2B compared to running a full LLM for routing decisions? AWS reports decisions in under 100 milliseconds on an Nvidia RTX 3090, with response times increasing as tasks grow more complex. On an M3 MacBook, the median latency for small tasks is around 150 milliseconds. This speed comes from restricting the model to selecting among supplied options rather than generating open-ended text. _Developers weighing latency tradeoffs for agent tool-calling can track benchmarks like this on daily.dev._ ### How does AWS Strands Decider 2B compare to OpenAI's Decisions API and TypeSafe's Jev model? Strands Decider is a downloadable, locally runnable model released with its training data and scripts, whereas OpenAI's Decisions API, powered by the Luna model, is a hosted limited preview. TypeSafe's Jev started the current wave of decision models; on JevBench's public set, Strands Decider ranks second among roughly 2-billion-parameter public models and first among those with a full published training recipe. _Teams choosing between hosted and local decision-model options can compare releases like these on daily.dev._ --- Tags: [#aws](https://daily.dev/tags/aws), [#ai-agents](https://daily.dev/tags/ai-agents), [#qwen](https://daily.dev/tags/qwen), [#lora](https://daily.dev/tags/lora) [View this post on daily.dev](https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs) ```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":"AWS launches a local answer to TypeSafe’s Jev decision model","url":"https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs"},"datePublished":"2026-10-01T16:07:22.917Z","dateModified":"2026-10-06T21:45:00.800Z","description":"AWS released Strands Decider 2B, a small downloadable decision model built on Qwen3.5-2B that scores developer-supplied answer options instead of generating...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/501e481232a1483cd7e15751cd952c05?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/501e481232a1483cd7e15751cd952c05?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","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 New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"aws,ai-agents,qwen,lora","timeRequired":"PT4M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"AWS launches a local answer to TypeSafe’s Jev decision model"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/aws-launches-a-local-answer-to-typesafe-s-jev-decision-model-uqpprqnhs#faq","mainEntity":[{"@type":"Question","name":"What is AWS Strands Decider 2B and how does it work?","acceptedAnswer":{"@type":"Answer","text":"Strands Decider 2B is a small decision model AWS released that scores developer-supplied answer options or returns numerical scores instead of generating free-form text. It uses Qwen3.5-2B as its base model, replacing the language-model head with a pointer head of just over a million parameters, trained with a rank-16 LoRA adapter, letting agents check proposed actions before they run. Teams validating agent actions before execution can follow decision-model releases like this one on daily.dev."}},{"@type":"Question","name":"How fast is AWS Strands Decider 2B compared to running a full LLM for routing decisions?","acceptedAnswer":{"@type":"Answer","text":"AWS reports decisions in under 100 milliseconds on an Nvidia RTX 3090, with response times increasing as tasks grow more complex. On an M3 MacBook, the median latency for small tasks is around 150 milliseconds. This speed comes from restricting the model to selecting among supplied options rather than generating open-ended text. Developers weighing latency tradeoffs for agent tool-calling can track benchmarks like this on daily.dev."}},{"@type":"Question","name":"How does AWS Strands Decider 2B compare to OpenAI's Decisions API and TypeSafe's Jev model?","acceptedAnswer":{"@type":"Answer","text":"Strands Decider is a downloadable, locally runnable model released with its training data and scripts, whereas OpenAI's Decisions API, powered by the Luna model, is a hosted limited preview. TypeSafe's Jev started the current wave of decision models; on JevBench's public set, Strands Decider ranks second among roughly 2-billion-parameter public models and first among those with a full published training recipe. Teams choosing between hosted and local decision-model options can compare releases like these on daily.dev."}}]} ```