Skip to content

LABBench2: An Improved Benchmark for AI Systems Performing Biology Research

A benchmark of realistic biology-research tasks involving literature, figures, tables, databases, and bioinformatics files.

7 models

Higher is better.

  1. 1Gemini 4 ArgonGoogle88.8
  2. 2Gemini 3.8 FlashGoogle86.2
  3. 3Claude Opus 5Anthropic84.2
  4. 4GPT-5.6 SolOpenAI82.1
  5. 5Gemini 3.7 FlashGoogle82.1
  6. 6GPT-5.6 TerraOpenAI81.2
  7. 7Claude Sonnet 5Anthropic80.1
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Aggregate accuracy
Tasks
Nearly 1,900 biology-research tasks
Authors
Jon M. Laurent, Albert Bou, Michael Pieler, Conor Igoe, Alex Andonian, Siddharth Narayanan, James Braza, Alexandros Sanchez Vassopoulos, Jacob L. Steenwyk, Blake Lash, Andrew D. White, Samuel G. Rodriques (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.