Amazon Web Services (AWS) has released Strands Decider 2B, an open-source decision model inspired by TypeSafe’s Jev. The announcement came the same week that OpenAI disclosed a similar offering.
What the model does
Strands Decider 2B provides a high-speed, low-cost way to choose between predefined options and returns a calibrated confidence score with each decision. The model is fully open-sourced, available now, and small enough to be run locally.
Development background
The project began as a homebrew prototype by Marc Brooker, an Amazon distinguished engineer, who built his own take after seeing Jev. That initial project performed well enough to briefly reach the top spot on the Jevbench ranking for models of its size. AWS engineers then cleaned up the prototype and released it from Strands Labs, an internal group focused on tools and protocols for deploying AI agents.
Brooker says the need for such a tool emerged in conversations with AWS customers whose agentic workflows do not always require the capability or cost of a full-fledged large language model. “What originally piqued my interest in this class of models was that they make a perfect decider for a workflow step — ‘what is the next thing for me to do here, based on where I am?’” Brooker told TechCrunch. He added that the model offers customers “a workflow step that can be structured in a way that is more reliable, thanks to the confidence scores, thanks to the closed domain of answers, [and is] lower latency, potentially lower cost.”
Technical details
Like other decision models, Strands Decider is built on the “torso” of an LLM — in this case Qen3.5-2B — but instead of generating free text it outputs calibrated choices. TypeSafe named their model Jev after economist William Stanley Jevons, referencing the idea that falling costs of a technology can increase its demand.
Market reaction and challenges
Dozens of similar models have been produced by researchers since TypeSafe introduced Jev, indicating strong interest in decision-focused intelligences. That volume of new models also prompts questions about their long-term value.
Brooker suggests the main challenge will be optimizing the model’s fast decision-making without degrading its intelligence. “There is a very careful balance to be found where you want to push its performance on accuracy and calibration on these kinds of tasks, without degrading its performance on understanding different languages, on having the kind of knowledge it has, which is what makes it general purpose and interesting and useful,” he told TechCrunch.
He does not necessarily expect frontier labs to dominate the space, noting that in smaller markets the cost to build something interesting can be in the hundreds or thousands of dollars.
TypeSafe’s CEO and founder Diogo Almeida, meanwhile, urged caution. He acknowledged that many see a "gold rush" around these models but warned people may underestimate the difficulty of making models genuinely smart. “The current batch seems more like ML people wanting to implement a cool architecture than a team deeply dedicated to making intelligence useful,” Almeida told TechCrunch.
Why this matters
Strands Decider 2B illustrates growing demand for compact, fast, and inexpensive decision models that can be embedded into agentic workflows without invoking a full LLM for every step. It also shows how quickly open-source prototypes and internal experiments can be matured into broadly available tools when they meet a practical need and have low resource requirements for local deployment.
