OpenAI Platform quietly adds a Decisions API playground with Predicate, Choice and Score modes
testingcatalog · x · 2026-10-20
TestingCatalog spotted a new "Decisions" API playground on the OpenAI Platform, letting users test OpenAI's recently released decision model. Screenshots show three modes: Predicate (returns the probability a statement is true, 0-100%), Choice (routing decisions like support-ticket triage), and Score (ordered-severity ratings, e.g. bug triage). It's a productized push to turn structured judgment into a standalone API alongside generative tasks.
More from coding & agent
- Devs are stripping Claude from GitHub contributors to hide AI-assisted work — ssh4net · 2026-10-20
- The future of CRM isn't a CRM — it's a context engine orchestrating agents — matt_slotnick · 2026-10-20
- NVIDIA paper predicts post-training coding-agent performance from base models with ρ=0.964 — sudoraohacker · 2026-10-20
- OpenAI benchmark: GPT-6.1 Sol hits 97% accuracy migrating COBOL at $0.50 per trial — ChrisGPT · 2026-10-20
- Matt Pocock: my self-built personal wiki became critical business infrastructure in 3 months — _AustinCalvert_ · 2026-10-20
- 99% don't know 'record a skill': computer use agents' ceiling is nowhere near — alliekmiller · 2026-10-20