OpenAI Releases LifeSciBench to Evaluate AI on Real Research Tasks, Comprising 750 Expert-Written Items Across 7 Biology Fields
According to OpenAI's official announcement on June 20, the company released LifeSciBench, a new evaluation benchmark designed to assess AI systems on real-world scientific research tasks. The benchmark comprises 750 expert-written tasks spanning 7 research workflows and 7 biology domains, created b
GateNews·06-20 14:55
