Amazon Jev Clone Brings Fast Decision Models to Developers
Amazon has released its own Amazon Jev clone, an open source decision model designed for automation rather than conversation. The model, called Strands Decider 2B, arrived the same week OpenAI announced a similar offering, showing that major tech companies now see specialized decision-making tools as a key part of agentic AI workflows.
Unlike large language models that generate text, the Amazon Jev clone focuses on sorting between pre-decided options and delivering confidence scores for each choice. It is fully open sourced, available now, and small enough to run locally, making it accessible to developers who need fast, low-cost automation.
What Makes the Amazon Jev Clone Different
The Amazon Jev clone is a high-speed, low-cost tool built to handle one specific job: deciding what happens next in a workflow. According to Amazon distinguished engineer Marc Brooker, the project began after he saw Jev and tried building his own version. That homebrew effort was successful enough — it briefly reached the top spot on the Jevbench ranking for models of its size — that Amazon engineers cleaned it up and released it through Strands Labs, an organization developing new tools and protocols for deploying AI agents.
Brooker said the need for this kind of model emerged from conversations with AWS customers. Many agentic workflows, he explained, do not always require the capability or cost of a fully featured LLM.
“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 said. He described the offering as “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.”
How Decision Models Work
Like other decision models, the Amazon Jev clone is built on the “torso” of an LLM — in this case Qwen3.5-2B — but instead of generating text, it delivers calibrated choices. This architecture allows the model to provide a measure of how confident it is in its selection, which can make automated workflows more reliable.
TypeSafe named its model Jev after the economist William Stanley Jevons, invoking his theory that the falling cost of something — like computer intelligence — can actually increase its demand. Since TypeSafe debuted the idea, dozens of similar models have been produced by researchers, showing wide interest in the approach. That proliferation also raises questions about how valuable these models can ultimately be.
Brooker suggests the challenge lies in optimizing the model’s speedy decision-making without compromising 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 said.
Open Source and Local Deployment
One of the most notable aspects of the Amazon Jev clone is its accessibility. The model is fully open sourced and available now, and its small size means it can run locally rather than requiring cloud infrastructure. This could appeal to developers who want more control over their agentic workflows or who need to avoid the latency and cost associated with larger models.
Brooker said he does not necessarily expect frontier labs to dominate the decision model space. With smaller markets, the cost to build something interesting is in the hundreds or thousands of dollars, making it feasible for a wider range of researchers and developers to experiment.
TypeSafe Responds to Growing Interest
For their part, TypeSafe executives say they are keeping their heads down and improving future models. CEO and founder Diogo Almeida said that while outside interest is growing, making these models genuinely smart is harder than it looks.
“I get that people think it’s a gold rush, but they might be underestimating the difficulty of making the models actually smart,” Almeida said. He added that for now, he did not see real competition for his company emerging yet.
“The current batch seems more like ML people wanting to implement a cool architecture than a team deeply dedicated to making intelligence useful.”
What This Means for AI Development
The release of the Amazon Jev clone highlights a broader shift in AI development. While frontier LLMs continue to push the boundaries of what is possible with text generation and reasoning, a parallel track is emerging around smaller, specialized models that excel at narrow tasks. Decision models like Strands Decider 2B are designed to be reliable, fast, and cheap — qualities that matter for businesses building agentic workflows at scale.
Whether these models will prove as valuable as their proponents hope remains to be seen. But with Amazon, OpenAI, and a growing number of researchers entering the space, the decision model category is clearly one to watch.
