OpenAI’s Jev clone could help the frontier lab stop its swarming agents
At OpenAI's Dev Day event on Tuesday, CEO Sam Altman announced a new "Decisions API," which he described as a way to give the company's Luna model a predefined set of options to choose between, such as categories for classifying an image or different agent behaviors.
According to Altman, focusing the model on that choice makes it extremely fast while keeping capabilities like image understanding, broad language support, and safety protections.
The API provides functionality similar to Jev, a model released by TypeSafe AI earlier this month that is explicitly designed for software automation. Jev is a classifier built on an LLM that developers can give a set of choices, which it outputs as probabilities cheaply and at high speeds.
TypeSafe did not respond to TechCrunch's questions about the new product. CEO Diogo Almeida, a former OpenAI engineer, joked on X about the beginning of the clone wars and said OpenAI's interest could be a sign 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 it applies to deliberate reasoning.
OpenAI released Decisions API as a limited preview, and TechCrunch said it has not yet spotted developers running it through its paces. According to conversations on X, there is clearly interest. Decisions API is not the only Jev-like API available; other startups are rolling out similar models.
Almeida told TechCrunch last week that his company's moat is the synthetic data it creates to generate statistically useful outputs, adding that fast and cheap is easy but intelligence is the hard part.
One likely application is monitoring and securing AI agents. Following a series of incidents where its agents misbehaved on the open internet, OpenAI introduced security measures including using a separate model to watch for bad actions at significant compute cost. Shapor Naghibzadeh, who leads the startup QueryStory, built a hackathon demo that uses Jev to check each agentic action against its given task, blocking actions it had high confidence were bad, flagging others for review, and permitting the rest. The report states such monitoring costs $2.94 with Jev versus $372 with a frontier LLM.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
OpenAI's "Decisions API" is a Jev clone that confirms the importance of fast, cheap intelligence.