Tsinghua and Partners Release ASI-Bench to Test AI Scientific Autonomy

Tsinghua, MIT, Harvard and other institutions released ASI-Bench, a benchmark of 60 real research projects across 11 scientific fields built by 40+ experts over 31,000 hours to measure AI scientific autonomy. Tests revealed that naming only a method for agents sharply degrades performance.

2026-08-25 ~ 2026-08-26 · 3 related posts

• Naming Methods Hurts Agents: Steps Drive Performance — rohanpaul_ai · 2026-08-25
• ASI-Bench evaluates AI's autonomous scientific research capabilities — rohanpaul_ai · 2026-08-25
• Tsinghua Releases ASI-Bench: 60 Tasks to Test AI's Scientific Autonomy — 机器之心 · 2026-08-26