OpenAI has made its Decisions API widely available, adding voice and visual support. Early customer Speak reported that the gap between actions fell to .13 seconds from 1.4 seconds.
OpenAI’s Decisions API uses a reconfigured version of ChatGPT Luna to speed up inference and lower costs, rather than a specially trained decision model. OpenAI also distributed inference resources geographically to reduce delays across regions.
TypeSafe’s Jev helped establish the category of AI models that quickly choose between options and report confidence in their decisions. In initial trials, engineers at Box, Datadog, Ramp and Speak used OpenAI’s API to replace LLMs with higher throughput.
For builders, decision models offer an alternative to LLMs for workflows centred on fast choices. The best fit still depends on the workload, with trade-offs between intelligence, cost, confidence calibration and speed.
General availability gives developers another supplier to test for fast decision-making workflows.
Voice and visual support extend the API’s potential uses beyond text classification.
Early customer speed gains do not prove which model will work best on a company’s own tasks.
Written automatically by the Made in 10 Newsroom from the source above.
Flapping Airplanes has raised a $1 billion Series A funding round led by Index Ventures and Lux Capital. The AI lab aims to reduce the data and computing power needed to train models, but has no product on sale yet.
DeepSeek has raised about $15 billion, TechCrunch reported, citing Bloomberg. The Chinese AI company is spending more on infrastructure and open models to serve customers in China and abroad.
Hugging Face has released Carbon-X, an open-source AI model that answers questions about DNA in natural language. The model can run locally on a laptop, keeping sensitive genetic data off cloud services, but its predictions lack wet-lab validation.