Today for AI
HOT RADAR
llmHEAT 31.1°

AI CURATED HOT STORY · 10/23/2026

OpenAI Launches Decisions API, Boosting Speed Tenfold

2 reports archived2 independent sourcesupdated 10/23/2026, 03:00:00
Synthesis & Latest Updates
2 Sources Cross-Validated

OpenAI has officially launched the Decisions API, built on the GPT-6 Luna architecture, to compete directly with TypeSafe’s Jev model. Designed for real-time micro-decisions with bounded answers, this new interface achieves up to 10x faster inference speeds compared to the standard Responses API across text, image, and structured inputs, significantly reducing latency and costs.

Technically, OpenAI adopted an agile strategy by reconfiguring existing ChatGPT Luna models rather than training from scratch. The API introduces native JSON support, prompt caching, and data residency options in the US and Europe, handling up to 400 questions per request. Acting as a 'super classifier,' it replaces complex LLM reasoning with rapid classification and confidence scoring in specific workflows.

This move signals a shift in AI infrastructure towards high-frequency, low-latency automated decision-making. By filling the gap in lightweight decision markets, the Decisions API offers developers a cost-effective solution for agent routing and content classification, accelerating the transition of AI applications from proof-of-concept to large-scale production.

LATEST/OpenAI has officially released the Decisions API, highlighting its capabilities in agent workflow classification and routing. With integrated native JSON support and data residency options, developers can now leverage its low-cost features to replace traditional LLMs for high-frequency micro-decisions, strengthening OpenAI's position in the real-time decision market.

HEAT TRENDHourly heat curve

2 fully observed hours
Fewer than 3 fully observed hours; trend not plotted yet.

TIMELINECoverage timeline

Total 2 reports · Latest first
  1. TechCrunch AIT2·78 pts
    • OpenAI launches Decisions API, directly challenging the fast, low-cost decision model market pioneered by TypeSafe's Jev.
    • These models act as 'super classifiers,' replacing traditional LLMs in specific workflows with lower cost and higher speed.
    • OpenAI adopted an agile strategy, reconfiguring existing ChatGPT Luna architecture instead of training a specialized model from scratch.
  2. OpenAI 官方博客T1·68 pts
    • Built on GPT-6 Luna, the Decisions API optimizes for bounded answers, delivering up to 10x faster decision-making than the Responses API.
    • New features include native JSON input support, up to 400 questions per request, prompt caching, and US/EU inference residency.
    • Designed for high-frequency micro-decisions in applications and agents, such as classification, routing, and next-action selection.