OpenAI launches Decisions API, challenging TypeSafe's Jev model
OpenAI has made its Decisions API widely available, entering the market for fast, low-cost decision models. This move competes with TypeSafe, a startup valued at $7.6 billion for its Jev model. By reconfiguring its Luna model, OpenAI reduces inference times to under 200 milliseconds, offering developers a cheaper alternative to large language models for classification tasks.
Key points
- OpenAI, the developer of ChatGPT, has made its Decisions API available to all customers, directly competing with TypeSafe, a startup founded by a former OpenAI researcher.
- TypeSafe recently reached a $7.6 billion valuation after its Jev model established a category for 'super classifiers' that make rapid decisions with confidence ratings.
- OpenAI’s product lead Nikunj Handa stated that the team reconfigured the existing ChatGPT Luna model to achieve inference speeds of 100 to 180 milliseconds, down from one to three seconds.
- The Decisions API, initially announced at OpenAI’s developer day on September 29, now includes new voice and visual modalities in its public release.
- Developers are adopting these 'System One' models to replace large language models in workflows requiring speed and cost efficiency, as they operate at a dramatically lower cost.
- While OpenAI currently uses a reconfigured Luna model, Handa indicated that specific training for decision-making capabilities may be developed in the future.
OpenAI has expanded the availability of its Decisions API, marking a significant entry into the emerging market for high-speed, low-cost decision-making models. This strategic move positions the company against TypeSafe, a startup that recently achieved a $7.6 billion valuation. TypeSafe’s Jev model pioneered the category of "super classifiers," which are deep learning systems designed to rapidly select options from a set and provide confidence ratings. The rapid adoption of these models has led developers to replace traditional large language models in various workflows, prioritizing speed and cost efficiency over general-purpose reasoning.
Nikunj Handa, the product lead for OpenAI’s APIs, explained that the team’s response to TypeSafe’s success was to reconfigure existing infrastructure rather than build a new model from scratch. The Decisions API utilizes the ChatGPT Luna model, modified to accelerate inference and reduce costs. Handa noted that this approach shrinks processing times from one to three seconds to between 100 and 180 milliseconds. Although the current implementation relies on the standard Luna architecture, OpenAI is considering specific training methods to enhance decision-making performance in future iterations.
The public release of the Decisions API, which was first announced at OpenAI’s developer day on September 29, now includes additional voice and visual modalities. This expansion allows developers to integrate fast decision logic into a broader range of applications. The competition highlights a shift in the AI industry toward specialized, efficient models for specific tasks, challenging the dominance of large, general-purpose language models in latency-sensitive environments.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.
