What Is the OpenAI Jev Clone?
The OpenAI Jev clone arrived quietly, tucked inside a larger announcement. At the company’s Dev Day event on Tuesday, CEO Sam Altman revealed a new “Decisions API” during an aside that quickly drew attention. The API appears to offer functionality similar to Jev, a model released by TypeSafe AI earlier this month that is explicitly designed for software automation. This OpenAI Jev clone signals that fast, cheap decision-making models are becoming a priority for the frontier lab.
Jev is a kind of super-powered classifier built on an LLM. Developers give it a set of choices, and it outputs probabilities cheaply and at high speeds. OpenAI’s Decisions API seems to be the same sort of product. Altman described the API as a way to give the lab’s Luna model a predefined set of options to choose between, such as categories in which to classify an image or different agent behaviors.
“By focusing the model on that choice, we can make it extremely fast while keeping capabilities like image understanding, broad language support, and safety protections,” Altman said at the event.
Why Fast Decision Models Matter
The subtext is that LLMs as we know them aren’t the right solution for a lot of software because they are comparatively slow and expensive. Developers have been using Jev to augment LLMs and have found that they’re faster and cheaper in the process.
It’s not clear how similar Decisions API will be to Jev, since OpenAI released it as a limited preview. So far, developers haven’t been spotted running it through its paces publicly. However, there is clearly interest, according to conversations on X.
TypeSafe didn’t respond to questions about the new product. But CEO Diogo Almeida, a former OpenAI engineer who co-invented reinforcement learning, joked on X about the beginning of the clone wars. He added that OpenAI’s interest could be “a sign…that building in a System One compatible way is the future.”
“System One” is TypeSafe’s term of art for fast, intuitive thinking, versus “System 2,” which it applies to deliberate reasoning.
How the OpenAI Jev Clone Could Secure AI Agents
After just weeks, it seems clear that these models have a future ahead of them. One likely application is monitoring and securing AI agents. This is especially relevant for OpenAI, which introduced new security measures following a series of incidents where its agents misbehaved on the open internet.
One of OpenAI’s new security measures involves using a separate model to watch for bad actions at “significant compute cost.” Shapor Naghibzadeh, a long-time cybersecurity professional who leads the startup QueryStory, thinks that a model like Jev could make that possible far more cheaply.
He built a demo for a hackathon held last weekend that uses Jev to check each agentic action against the task it was given. The system blocks actions it has high confidence are bad, flags others for review, and permits the rest.
The Cost Difference Is Staggering
In theory, such monitoring could have stopped the Hugging Face incident. Monitoring of that kind costs $2.94 with Jev, versus $372 with a frontier LLM. That price difference is staggering for anyone running agents at scale.
A key observation is that Jev is arguably cheap enough to run on every agentic action. This offers a layer of review that could improve the reliability of agents writ large. It’s the kind of thing TypeSafe was hoping to achieve, and now OpenAI has seen the value as well.
Decisions API isn’t the only Jev-like API on the internet. Other startups are rolling out similar models, and OpenAI won’t be the last tech giant to produce one. A key question is how well calibrated each of these decision models’ outputs will be to real life.
Almeida says his company’s moat is the synthetic data it creates to generate statistically useful outputs.
“Fast and cheap is very easy, you know,” Almeida said last week. “If you want it really fast and cheap, use dice, right? Intelligence is the hard part, and my North Star is always pushing the intelligence-per-dollar Pareto curve.”
What the OpenAI Jev Clone Means for Developers
For developers building with AI agents, the arrival of the OpenAI Jev clone is significant. It means a major lab is validating the idea that not every task needs a full frontier model. Sometimes you just need a fast, cheap classifier that can make reliable decisions.
The clone wars Almeida joked about may be just beginning. As more companies release decision models, the competition will likely focus on calibration and intelligence-per-dollar rather than raw speed alone. Developers should watch how these models perform in real-world agent monitoring, where the stakes are high and the margin for error is small.
Monitoring Every Agentic Action
The ability to run a decision model on every single agentic action changes the economics of AI safety. Instead of sampling actions or relying on expensive frontier models, teams can check each step an agent takes. That kind of continuous review could catch misbehavior before it causes real harm.
OpenAI’s own agents have already shown why this matters. A series of incidents where agents misbehaved on the open internet pushed the lab to add new security measures. Those measures currently come at significant compute cost, which is exactly the problem a cheaper decision model could solve.
For now, OpenAI’s Decisions API remains in limited preview. But its existence, and the interest it has generated, suggest that the OpenAI Jev clone is more than a passing experiment. It could become a foundational tool for keeping swarming AI agents in check.
