That’s… not how thinking models work. They tend to be iterative and serial, not ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		brookst 6 months ago \| parent \| context \| favorite \| on: AI agent benchmarks are broken That’s… not how thinking models work. They tend to be iterative and serial, not parallel and then pick-one.

e1g 6 months ago [–]

Parallel test time compute is exactly what SOTA models do, including Claude 4 Opus extended, o3 Pro, Grok 4 Heavy, and Gemini 2.5 Pro.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact