The AIntibody Challenges
AIntibody is a series of benchmarking exercises that aims to advance antibody discovery by assessing the ability of unique AI algorithms to enhance or expedite the process. As the exercises—or ‘competitions’—progress, each challenge will become progressively more difficult.
Results have now been published in Nature Biotech here: https://www.nature.com/articles/s41587-026-03238-6
Across 511 submissions from 29 organizations, the results are encouraging but nuanced. Several groups generated high-affinity, developable antibodies, including strong performance in affinity maturation. At the same time, success did not consistently transfer across tasks, and some approaches struggled to outperform simpler experimental or statistical baselines.
The takeaway: AI can already contribute meaningfully in defined, data-rich antibody discovery workflows, but broad, reliable generalization remains an important challenge for the field.
Preliminary insights from the AIntibody Benchmarking Challenge
View our webinar to see the early findings from 33 participating teams.
The Challenges – Find out more about the 3 challenges