Development

A biology model raises the pass rate on 750 expert tasks

development2026-06-21

Claim

OpenAI unveiled LifeSciBench, 750 expert biology tasks, where GPT-Rosalind raised the pass rate to 36.1% from GPT-5.5's 25.7%, strong on writing yet weak on design.

Reported metric: 36.1% vs 25.7%

Reporting date: 2026-06-21 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to June 21, 2026 (theinnermostloop.substack.com)

Knowledge graphopen in browser →

Reported in

Welcome to June 21, 2026

Actor

OpenAI

Evidence for

Supersedes

A near-autonomous AI chemist runs 10,080 experiments and finds the trick

Referenced by

Welcome to June 21, 2026

Cite or correct this record

Citation formatCitation text

A biology model raises the pass rate on 750 expert tasks
Innermost Loop KB — secondary record
Reporting date: 2026-06-21
KB permalink: https://nicholsn.github.io/innermost-loop-kb/developments/2026-06-21-a-biology-benchmark-of-750-expert-tasks
Claim: OpenAI unveiled LifeSciBench, 750 expert biology tasks, where GPT-Rosalind raised the pass rate to 36.1% from GPT-5.5's 25.7%, strong on writing yet weak on design.
Attributed sources:
- Welcome to June 21, 2026 — https://theinnermostloop.substack.com/p/welcome-to-june-21-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.