One of the most compelling announcements made during OpenAI’s Dev Day event on Tuesday came from CEO Sam Altman, who revealed the company’s new “Decisions API.”
This API appears to offer functionality comparable to Jev, a model released by TypeSafe AI earlier this month that is specifically built for software automation. Acting as a super-powered classifier based on an LLM, Jev allows developers to input a set of choices, which the model outputs as probabilities, doing so cheaply and at high speeds.
OpenAI’s Decisions API seems to serve a similar purpose. At the event, Altman described the API as a mechanism to provide the lab’s Luna model with a predefined set of options to select from, such as image classification categories or specific agent behaviors.
“By focusing the model on that specific choice, we can make it extremely fast while retaining capabilities like image understanding, broad language support, and safety protections,” Altman stated.
TypeSafe did not respond to TechCrunch’s inquiries regarding the new product, but CEO Diogo Almeida, a former OpenAI engineer who co-invented reinforcement learning, jokingly referenced the “clone wars” on X.
He noted that OpenAI’s interest might signal that “building in a System One compatible way is the future.” (System One is TypeSafe’s term for fast, intuitive thinking, as opposed to “System 2,” which they apply to deliberate reasoning.)
The underlying message is that current LLMs are not the optimal solution for many software tasks due to their comparative slowness and expense. Developers have utilized Jev to augment LLMs, discovering that it offers significantly faster and cheaper performance.
It remains unclear how similar the Decisions API will be to Jev, as OpenAI released it as a limited preview, and TechCrunch has not yet observed developers testing it extensively. However, there is clearly significant interest, as evidenced by the discussions on X.
The Decisions API is not the only Jev-like API available on the internet, as other startups are rolling out similar models, and OpenAI will likely not be the last tech giant to do so. A key question remains regarding how well calibrated the outputs of these decision models will be in real-world scenarios.
Almeida asserts that his company's competitive advantage lies in the synthetic data it creates to generate statistically useful outputs.
“Fast and cheap is very easy, you know,” Almeida told TechCrunch 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.”
After just a few weeks, it appears clear that these models have a significant future, with one likely application being the monitoring and securing of AI agents. One of OpenAI’s new security measures, implemented following a series of incidents where its agents misbehaved on the open internet, involves using a separate model to detect bad actions at “significant compute cost.”
Shapor Naghibzadeh, a long-time cybersecurity professional leading the startup QueryStory, believes that a model like Jev could make this monitoring far cheaper.
He built a demo for a hackathon held last weekend that utilizes Jev to check each agentic action against the assigned task, blocking actions with high confidence of being malicious, flagging others for review, and permitting the rest.
In theory, such monitoring could have prevented the Hugging Face incident, with monitoring of this type costing only $2.94 using Jev compared to $372 using a frontier LLM.
A key observation is that Jev is arguably cheap enough to run on every agentic action, providing a layer of review that could significantly improve the reliability of agents on a large scale. This represents the kind of outcome TypeSafe hoped to achieve, and it appears OpenAI has now recognized the value as well.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.