All Competitors

Transformer framework that models protein-protein interactions at residue resolution, generalizing zero-shot to unseen MHC alleles and sequence-neutral PTMs from one fixed checkpoint.

SMILES-based chemical language models pretrained on 100M+ molecules to natively represent therapeutic peptide chemistry, including non-canonical residues.

A 730M-parameter protein foundation model that co-designs enzyme sequence and 3D structure under small-molecule ligand guidance for de novo enzyme design.

Protein language model that predicts RNA-binding domains, global RNA-binding propensity, and mutation effects at single-residue resolution from sequence.

A DeepSeek-7B-based multi-task large reasoning model that applies chain-of-thought reasoning and reinforcement learning across ~10 molecular science task families.

Foundation-model-guided virtual staining that generates four IHC markers (HER2, Ki67, ER, PR) from H&E using a single SPADE-UNet conditioned on a frozen UNI encoder.

A 2.6B-parameter Liquid Foundation Model fine-tuned via Insilico's MMAI Gym to handle ADMET, retrosynthesis, drug-target activity, and molecular optimization in one checkpoint.

A multimodal, domain-constrained foundation model that self-supervises on ~500K PubChem molecules to jointly predict nine thermophysical properties of small molecules.

Multimodal deep learning model that predicts protein-mediated chromatin contact maps and loops de novo from protein-binding profiles and sequence across cell types.

An all-in-one foundation model for microscopic image restoration, unifying 8 tasks across 5 microscopy modalities and 2D/3D data, with zero-shot inference on unseen imaging systems.

A pretrained multi-task flow-matching model for structure-based drug design, unifying de novo design, docking, conformer generation, and pharmacophore conditioning.

A 1.2B-parameter fMRI foundation model of brain connectomes that uses brain-environment interaction tokens and multitask learning for behavior and disease prediction.

A self-supervised foundation model for structural brain MRI combining DenseNet and Vision Transformer with Mixture of Experts for multi-task brain disease diagnosis and brain age prediction.

A modality-agnostic, multi-task foundation model for human brain imaging that runs five core tasks across uncalibrated CT and MRI without retraining.

Multimodal multitask foundation model that fuses 3D low-dose chest CT with clinical data to perform 17 lung-cancer-screening tasks via free-text question answering.

Biomedical multimodal LLM with pixel-level insight, combining visual question answering, pixel-grounded prompts, and segmentation via a mixture-of-experts design.

Multi-modal, multi-task biological foundation model trained on 2 billion samples spanning proteins, small molecules, and single-cell gene expression.

A general CT image segmentation foundation model that uses task-prompted automatic pathway decoding to segment 83 anatomical structures and lesions across whole-body CT.

EEG-to-language multi-task foundation model that pairs a Q-Conformer encoder with frozen LLMs to decode coherent sentences from non-invasive brain signals.

Deep learning model predicting single-cell read counts from DNA sequence features and cell transition graphs to identify transcriptional regulators.

Open-source, lightweight generalist vision-language foundation model for diverse biomedical imaging and text tasks.

A unified multi-task framework that converts diverse protein prediction tasks into autoregressive next-token prediction using pre-trained protein language model encoders.

Self-supervised CNN pretrained on 700,000 person-days of UK Biobank accelerometer data for human activity recognition, transferable across devices and cohorts.

A 7B-parameter protein language model built on LLaMA-2 that performs both protein sequence generation and superfamily classification in a unified framework.