Skip to content
TechCrunch·

🤖OpenAI Unveils Decisions API Model for Cheaper, Faster Intelligence

OpenAI's Decisions API is here, and it's a game changer for cost-conscious teams

TL;DR

OpenAI released its Decisions API model, offering faster and cheaper intelligence compared to traditional LLMs. It's now available to all customers and includes new voice and visual modalities. Developers can expect significant performance improvements in their workflows.

OpenAI just made its Decisions API model widely available, promising faster and cheaper intelligence compared to traditional large language models (LLMs). This is a big deal for developers looking to optimize costs and performance in their workflows. The Decisions API model operates at a dramatically lower cost and can cut the time between actions from 1.4 seconds to 0.13 seconds, making it a game changer for applications requiring quick decision-making. OpenAI also added new voice and visual modalities to the public version, ensuring lower latency in regions around the world by targeting its inference resources geographically. The model was announced on September 29 and became available to all customers on Thursday.

OpenAI Unveils Decisions API Model for Cheaper, Faster Intelligence — TechCrunch

Key Points

1

Decisions API model announced on September 29, available to all customers on Thursday

2

Operates up to 10x faster and at a dramatically lower cost compared to traditional LLMs

3

Cuts time between actions from 1.4 seconds to 0.13 seconds in initial trials

4

Includes new voice and visual modalities, ensuring lower latency in regions worldwide

5

TypeSafe's Jev model ranks seventh and third on public benchmarks, competing with D1 Ultra

Why It Matters

If you're running decision-heavy workflows that currently rely on LLMs, the Decisions API model could slash costs and improve performance. For instance, initial trials with customers showed a 1.4 second to 0.13 second reduction in action time, making it a no-brainer for applications requiring quick, high-stakes decisions. However, the model's performance on public benchmarks like JevMan and D1 Ultra varies, so it's worth testing in your specific use case.

OpenAIDecisions APILLMscost optimizationperformance improvement

Comments

Subscribe to join the conversation...

Be the first to comment

Enjoyed this article?

Get it daily. 7am. Free. Reads in 5 minutes.

Join 3,569 builders reading daily.