All Competitors

Family of autoregressive genomic foundation models that reconcile k-mer tokenization with single-nucleotide resolution at contexts up to 98k bp.

Single-cell foundation model using tabular attention over context cells to predict responses to arbitrary perturbations without fine-tuning.

Antibody developability predictor pairing text and protein language models, using in-context learning to fit new assays without retraining.

In-context learning model for cell-free RNA, meta-trained on synthetic tasks from a cfRNA structural causal model for few-shot cancer classification.

Graph diffusion transformer for in-context molecular design, adapting to new tasks from a few molecule-property demonstrations without fine-tuning.

Tabular foundation model adapted for extreme feature counts, enabling in-context prediction on wide omics tables with tens of thousands of features.

Protein function prediction via compressed in-context learning on a sequence-structure language model, cutting 751-token demonstrations to under 16.

Protein foundation model built on Bayesian Flow Networks, prompted with MSA profiles for structure- and function-preserving sequence design.

Genomic prediction model for plant and animal breeding, pretrained entirely on simulated populations and deployed with no training or tuning.

Epitope-conditioned T cell receptor generator that writes its own in-context examples, so receptors can be designed for targets with no known binders.

Chemical language model generating SMILES on a recurrent xLSTM backbone, designing within an unseen molecular domain from a few in-context examples.

Homology-aware protein language model on a recurrent xLSTM backbone, generating and scoring sequences from long contexts of unaligned homologs.

RNA folding kinetics model that predicts the full distribution of first passage times from a few simulated examples in a single forward pass.

Mixed-modal DNA, RNA, and protein foundation model at 110M and 270M parameters, with in-context learning across sequence modalities.

Protein fitness prediction model meta-trained across ProteinGym deep mutational scans, transferring to unseen assays in context with no task data.

Multimodal medical vision-language model for few-shot visual question answering, learning new imaging tasks from in-context examples at inference.

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