All Competitors

Single-cell foundation model pretrained on about 30 million human transcriptomes, using rank-value encoding for context-aware gene network inference.

Self-supervised XCiT encoders for prostate histopathology, pretrained on 48 million tissue tiles so that features cluster by histological pattern.

All-atom protein diffusion model that co-designs backbone, sequence and sidechains by denoising a superposition over all 20 sidechain states.

Protein structure prediction for monomers and complexes in one three-track network, scaling past 1000 residues without triangle attention.

Multimodal pathology assistant that answers questions about histology and cytology images, pairing the PathCLIP vision encoder with a Vicuna-13B LLM.

Spatial proteomics imputation from a 7-plex immunofluorescence panel, generating in silico CODEX expression for 33 more biomarkers per cell.

Protein model quality assessment predicting per-residue lDDT for monomer and multimer interface models from graph-coupled ESM embeddings.

Multi-modal LLM answering free-form questions about a compound's indications, pharmacodynamics and mechanism of action from its SMILES string.

Vision-language framework for 3D medical image diagnosis and visual question answering, bridging frozen image encoders and LLMs, shown on brain MRI.

Generative medical visual question answering model that pairs a vision encoder with a language model, trained on the 227k-pair PMC-VQA dataset.

Multi-domain protein and complex assembly from deep-learned inter-domain interactions, averaging TM-score 0.922 across 219 multi-domain targets.

Zero-shot antibody affinity maturation using ESM pseudolikelihood scoring. Improves binding up to 160-fold with no antigen-specific training data.

Single-cell foundation model pre-trained on 22 million transcriptomes, using rank-based gene encoding for clustering and trajectory inference.

Drug pair synergy prediction for rare cancer tissues, read from a language model's representation of a screening row written out as a sentence.

Scalable and transferable U-Net family (14M–1.4B parameters) for 3D medical image segmentation, supervised-pretrained on TotalSegmentator.

Medical image segmentation model that solves unseen segmentation tasks in context from a few labeled examples, with no retraining or fine-tuning.

Rigid protein-protein docking by diffusion over the rigid-body pose, with a confidence model ranking sampled complexes. Median C-RMSD 4.85 on DIPS.

Protein structure encoder pretrained by contrastive alignment to a frozen protein language model, anchored by self-supervised contact-map prediction.

Biomedical vision-language model trained contrastively on 1.6M figure-caption pairs mined from PubMed Central open-access articles.

Joint sequence-structure protein representation framework that fuses ESM-2 language model embeddings with GearNet geometric graph neural networks.

Structure-based drug design by SE(3)-equivariant diffusion over 3D atom coordinates and types, with the same frozen network scoring binding affinity.

Remote-homolog template recognition that threads a sequence against clustered PDB and AlphaFold DB structures to improve AlphaFold2 modelling.

Biomedical vision-language model trained contrastively on 15M PubMed Central figure-caption pairs for zero-shot classification, retrieval, and VQA.