All Competitors

5.6M-parameter multimodal foundation model fusing fMRI time series with diffusion-MRI structural connectivity in a shared ROI embedding space.

Protein inter-residue distance prediction fusing MSA Transformer coevolution with ESM2 sequence features, reaching a mean absolute error of 2.20 Å.

Transformer pretrained on raw fNIRS brain signals that scores procedural skill and transfers to unseen surgical procedures via a tiny adapter.

CLIP-based vision-language foundation model for eye imaging, enabling zero-shot disease detection and cross-modal retrieval across 11 modalities.

Tandem mass spectrometry model that embeds MS/MS spectra and molecular graphs in one space, ranking candidate structures without a spectral library.

Medical multimodal LLM (2B and 8B) trained for generalizable, step-by-step clinical reasoning via Mentor-Intern Collaborative Search.

Catalytic residue prediction from enzyme sequence alone, pairing frozen ESM2 embeddings with contrastive learning. Reaches F1 0.86 on Uni3175.

Structure-free ligand generation conditioned on a protein sequence alone, using masked diffusion over SMILES trained on 1.2M BindingDB active pairs.

Multimodal drug-response model coupling cell and molecule foundation models, pretrained on 1.8M perturbation RNA-seq profiles over 22,000 compounds.

Peptide-HLA immunogenicity prediction with a BiLSTM ensemble, inside a pipeline that finds microbial epitopes mimicking tumor neoantigens.

Molecular property prediction foundation model pretrained on computed descriptors. Wins 75% of Polaris tasks and 97% of MoleculeACE assays.

Bulk transcriptome foundation model, 150M parameters over ~20,000 protein-coding genes. Imputes masked expression at Pearson r = 0.954.

Structure-based drug design model pairing SE(3)-equivariant diffusion with retrieval of pocket-matched scaffolds to generate ligands for a target.

Text-guided protein editing framework with disentangled structure and function latents, edited by rewriting either description at inference time.

MSA design model generating alignments from protein language model embeddings to improve folding accuracy on orphan and low-homology proteins.

Graph-level self-supervised pretraining for 3D molecules, reconstructing whole-molecule geometry to sharpen quantum property and force prediction.

De novo peptide sequencing from tandem mass spectra, using curriculum learning and iterative self-refinement to stabilize non-autoregressive decoding.

Multimodal slide-level pathology foundation model trained by clinical-dialogue supervision on 2.3M whole-slide images and 14M Q&A pairs.

Antibody language model pretrained on 402 million OAS sequences, matching far larger antibody LMs on repertoire tasks at 125M parameters.

Protein-protein binding affinity and interface hotspot prediction from sequence alone, using protein language models fine-tuned on SKEMPI 2.0.

Contrastive language-image model for fMRI functional decoding, predicting cognitive tasks, concepts, and domains from brain activation maps.

Discrete diffusion model for protein sequence design in an all-atom SELFIES representation, reaching non-canonical and modified amino acid residues.

Multimodal viral foundation model over nucleotide and protein sequence, built for virus discovery, function annotation, and antibody design.