OpenAI Makes Decisions API Model Available for General Use Focused on Fast Decision-Making
OpenAI has made Decisions API, an optimized version of the ChatGPT Luna model, available to all customers. Standing out with its low latency, this new service aims to accelerate rapid decision-making processes.
OpenAI announced that it has rolled out the Decisions API model for general availability to all customers following a developer preview period. Designed as a re-architected version of the ChatGPT Luna model, this API was specifically built for decision-making tasks requiring high speed and low cost.
According to data shared by OpenAI API product lead Nikunj Handa, the new model reduces inference latency from one to three seconds down to 100-180 milliseconds. Announced during the developer day on September 29, the service now includes new audio and visual modalities alongside optimized geographic latency features.
Early adopters such as Box, Datadog, Ramp, and Speak reported increased efficiency driven by the model. Speak, in particular, noted that action latency dropped from 1.4 seconds to 0.13 seconds. With this move, OpenAI has entered into direct competition with startups like TypeSafe, which stands out in System One decision-making processes with its 'Jev' model category.
Sources
- techcrunch.com · Link