OpenAI has made its Decisions API model widely available, marking its entry into the competitive market for fast and cost-effective intelligence solutions. This move comes shortly after the rise of TypeSafe, a startup founded by a former researcher at OpenAI, which recently achieved a valuation of $7.6 billion with its Jev model. The Jev model has pioneered a new category of deep learning models known as super classifiers, which can quickly make decisions from a set of options and provide a confidence rating.

Context

Software developers have discovered that these decision models, also referred to as System One models, can effectively replace traditional large language models (LLMs) in various workflows, offering enhanced speed and significantly lower costs. Nikunj Handa, the product lead for OpenAI’s APIs, noted that upon seeing the Jev model, his team recognized numerous potential use cases, stating, "wow, we have like a million use cases for this."

Developments

In response, Handa’s team rapidly developed their own version of the decision model. Instead of creating a specifically calibrated model, they reconfigured ChatGPT Luna to enhance inference speed and reduce costs. Handa explained, "The big unlock with Decisions API has been the speed, shrinking things from one to three seconds to 100, 120, 180 milliseconds."

The Decisions API was announced during the company’s developer day on September 29 and became accessible to all customers on Thursday. OpenAI has also introduced new voice and visual modalities to the public version, aiming to lower latency globally by optimizing inference resources geographically.