IBM MAMMAL tops biology benchmarks

This is a meaningful multimodal biology paper, not just another protein model claim. The interesting part is the scope: MAMMAL is trying to unify sequence, molecule, and transcriptomics tasks that usually live in separate model families.

• –The benchmark spread matters more than the headline number: DTI, PPI, cell type annotation, cancer drug response, and antibody design sit at different points in the drug-discovery stack
• –The AlphaFold 3 comparison is directionally useful, but it is not a clean replacement story; the paper frames them as complementary tools for different biological questions
• –Public weights make this more than a paper win, since others can now test whether the reported gains hold outside the paper’s curated benchmarks
• –If the results generalize, the bigger shift is toward shared biomedical representations rather than one-off models for structure, affinity, or expression separately