GPT-6 Astra Scores 80.6 on OpenAI's New 'LifeVal' Benchmark, Which Even Grades Kindergarten Selection

shiritomo · 11/03 08:25 UTC

OpenAI has released a new evaluation benchmark that has AI handle everyday life tasks and scores the results. Examples include choosing a suitable kindergarten after a move and arranging volunteers in preparation for a flood. The latest model, GPT-6 Astra, is the one featured.

Read the original (shiritomo)

See it in the digest → ・ See this issue →