Trends

Research-agent evaluation shifts to harder domain-specific benchmarks

AI in scientific discovery · May 1 – May 1

Branched fromAI agents begin automating alignment research

Organizations

BAAI

Updates

  1. May 1New

    BAAI's AutoResearchBench found the best models reach only 9.39% accuracy tracing a target paper and 9.31% IoU collecting qualifying papers, far below their general browsing scores, showing generic agent benchmarks overstate research ability.

    BAAI's AutoResearchBench

Predictions

  1. Nov 2026

    At least two frontier labs name evaluation infrastructure for research agents (tamper-resistant benchmarks, live leaderboards, and automated grading for scientific tasks) as a stated strategic priority. May 1