OpenAI unveils Decisions API for faster, cheaper software automation choices
OpenAI announced a "Decisions API" at its Dev Day event on Tuesday that performs similar functions to Jev, a recently released model by TypeSafe AI designed for software automation. The API allows developers to give models a predefined set of options to classify or choose between, making it faster and cheaper than traditional large language models whilst maintaining capabilities like image understanding and safety protections. This announcement reflects growing demand for more efficient alternatives to standard LLMs for software automation tasks.
Jev operates as a specialised classifier that outputs probabilities quickly and cost-effectively, and OpenAI's Decisions API appears to follow a similar approach. TypeSafe AI's CEO, former OpenAI engineer Diogo Almeida, remarked humorously about the "beginning of the clone wars" whilst suggesting that this indicates fast, intuitive "System One" thinking represents the future of AI development. Beyond OpenAI, multiple startups are developing comparable decision-making models, though their effectiveness depends heavily on how well-calibrated their outputs are to real-world scenarios, with TypeSafe claiming its competitive advantage lies in synthetic data generation for statistically reliable results.
- OpenAI launches Decisions API, mirroring TypeSafe's Jev for faster, cheaper automation
- Fast, intuitive decision-making may be replacing slow, expensive traditional LLMs
- AI agent monitoring and security applications emerging as key use case
Read the full article at the source →
Originally published by TechCrunch as “OpenAI’s Jev clone could help the frontier lab stop its swarming agents”.