Papers & Research

LifeSciBench released

Original title:Introducing LifeSciBench

AI summary

In June 2026, OpenAI, together with 173 PhD-level life scientists, released the LifeSciBench evaluation benchmark, covering 750 real research tasks across seven workflows, including evidence handling, analysis and design optimization, and seven biological domains. Each task has about 25 detailed scoring criteria (19,020 in total) to assess scientific correctness and practical value. 79% of tasks require multi-step reasoning, and 53% require interpreting attachments such as charts and PDFs, aiming to measure AI's practical capabilities in complex, uncertain research tasks rather than only answering structured questions.

Key Takeaways

  • In June 2026,OpenAI,together with 173 PhD-level life scientists,released the LifeSciBench…
  • Each task has about 25 detailed scoring criteria (19,020 in total) to assess scientific…
  • 79% of tasks require multi-step reasoning,and 53% require interpreting attachments such as charts…

Quick Facts

Published
Updated
Source
OpenAI:官网动态(RSS · 排除企业/客户案例)
Category
Papers & Research
AI score
58
Read the original Read the full article at the original source
Why it matters

OpenAI 这个基准请了 173 位博士级科学家出题,第一次把 AI 评估拉到真实科研决策里。结果很实在:前沿模型在需要结合复杂图表、设计实验的任务上仍然乏力,做 AI for Science 的团队值得拿来校准预期。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: LifeSciBench released. June 17, 2026. https://new.nat6.net/en/news/cmqijfbdz01grsl5wxf4rd7e6/

View all →