Skip to content
AI News HubLIVE
More
Summary only

Task-Completion Time Horizons of Frontier AI Models

METR publishes updated measurements of AI agents' task-completion time horizons, showing exponential growth since 2019. As of May 2026, frontier models achieve about a 2-hour 50% time horizon on diverse software tasks. The article details methodology, limitations, and FAQs.

SourceHacker News AIAuthor: nsoonhui
Task-Completion Time Horizons of Frontier AI Models
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions

Summary only

Under the source content permission, this page provides an editorial summary, key points, and a link to the original article.

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Time horizon measures the longest task (by human expert time) an AI can complete with a given success probability.
  • Latest data shows exponential growth: 50% time horizon reached ~2 hours in 2026, up from seconds in 2019.
  • Evaluation uses over 100 software, ML, and cybersecurity tasks, but AI performance is uneven across domains.
  • Time horizon does not imply autonomous operation; tasks are low-context and not representative of real jobs.

Highlights and analysis are generated automatically and may contain errors. Check the original source.

More recent stories on this topic · Models