Skip to content

Anthropic protein-design library-ranking task

Ranks held-out protein designs for experimental testing using wet-lab campaign measurements and a selection objective.

2 models

Higher is better.

  1. 1Claude Opus 5.5Anthropic56.0
  2. 2Claude Sonnet 5.5Anthropic54.8
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Chance-normalized average precision
Tasks
83 protein-design library-ranking problems
Authors
Anthropic (2026)

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