Benchmark

METR time horizon

Also known as: METR

2stories this week
3last 30 days
3all time

Timeline

  1. Oct 8, 2026 · Opinion / analysis · 1 source
    [AINews] not much happened today
    More AI safety intrigue in the alignment below.
  2. Oct 8, 2026 · Research paper · 1 source
    On the estimation and validity of AI time horizons---a statistical look at the METR plot
    On 228 tasks and 26 AIs, we recompute the time horizons using splines and item-response theory to relax the assumption that the AI difficulty of a task depends linearly on the log of human time.
  3. Sep 30, 2026 · Model release · 1 source
    Last Week in AI #345 - 5 new models, 9 misalignment incidents, some Dots
    Anthropic and OpenAI race to release smarter and cheaper models

Often appears with