Structure & co-folding (24)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| AF2-Multimer Multi-chain fine-tune of AF2; the classic protein-complex baseline | Google DeepMind | Open | Public data only | 1 |
| AlphaFold 3 Biomolecular co-folding model that later co-folding models use as the reference baseline | Google DeepMind, Isomorphic Labs | Partly open | Public data only | 6 |
| BioEmu-2 All-atom diffusion model sampling structural ensembles of complexes and nucleic acids | Microsoft Research AI for Science | Open | Public data only | 0 |
| Boltz-1 / Boltz-1x First fully open co-folding model approaching AF3, with 1x adding physical-validity steering | MIT, Boltz PBC | Open | Public data only | 2 |
| Boltz-2 Open model that adds a small-molecule binding-affinity head on top of co-folding | MIT, Recursion, Boltz PBC | Open | Public data only | 3 |
| Chai-1 Open-weights co-folding model that stays stable without MSAs and accepts experimental restraints | Chai Discovery | Open | Public data only | 5 |
| Chai-2 Closed zero-shot antibody and binder design model with an upgraded folding module built in | Chai Discovery | Closed | Public data + wet lab for validation only | 0 |
| DiffDock Diffusion-based blind docking, now used mainly as the deep-learning docking baseline | MIT CSAIL | Open | Public data only | 2 |
| ESMFold2 Protein-language-model co-folding that needs no MSA retrieval, released alongside a sequence atlas | Biohub, EvolutionaryScale team | Unclear | Public data + wet lab for validation only | 2 |
| HelixFold-S1 A contact-guided sampling method added on top of HelixFold3, not a new base model | Baidu | Closed | Public data only | 0 |
| IntelliFold-2 Hong Kong co-folding model billed as an AF3 improvement; Pro tier is server-only | IntelliGen AI | Open | Public data only | 2 |
| IsoDDE Isomorphic Labs closed drug design engine extending AlphaFold 3 co-folding and affinity | Isomorphic Labs | Closed | Not disclosed | 2 |
| MassiveFold Large-scale sampling and parallelisation around AF2, not a new model; hard interfaces gain most | Lille University UGSF, CNRS | Open | Public data only | 2 |
| NeuralPLexer3 Flow-matching co-folding model emphasising stereochemistry and physical validity | Iambic Therapeutics | Closed | Public data only | 3 |
| OpenComplex2 All-atom generative model that samples conformational distributions of proteins, nucleic acids and complexes by graph diffusion rather than predicting one structure | Beijing Academy of Artificial Intelligence | Unclear | Not disclosed | 1 |
| OpenDDE Fully open 655M-parameter co-folding model tuned for antibody-antigen, billed as an open discovery engine | Aureka AI Research, Aureka Biotechnologies | Open | Public data only | 1 |
| OpenFold3 Consortium-built Apache-2.0 open reproduction of AlphaFold 3 with a federated private-data fine-tuning track | OpenFold Consortium, Open Molecular Software Foundation, Columbia University, Lawrence Livermore National Laboratory, Seoul National University | Open | Private or licensed data | 3 |
| Pearl Genesis Molecular AI's closed protein-ligand co-folding model, trained on large-scale physics-based synthetic structures | Genesis Molecular AI | Closed | Public data + wet lab for validation only | 2 |
| Protenix Open co-folding series reproducing and trying to exceed the contemporary closed model, with v2 weights turned proprietary | ByteDance Seed, ByteDance Seed AI for Science | Partly open | Public data + wet lab for validation only | 4 |
| RF3 Open co-folding model from a protein design institute, released under BSD with the AtomWorks data framework | University of Washington Institute for Protein Design, RosettaCommons | Open | Public data only | 1 |
| RhoFold+/DRfold2/trRosettaRNA Three independent RNA 3D prediction lines from three academic groups, not mutually comparable | CUHK / Harvard / MIT collaboration, National University of Singapore group, Shandong University (Qingdao) group | Partly open | Public data only | 2 |
| RNAPro Open model predicting all-atom RNA 3D structure, about 488M parameters | NVIDIA, Stanford University | Open | Public data only | 1 |
| SeedFold ByteDance Seed AF3-style co-folding model that scales by widening the Pairformer, with a linear triangle-attention variant | ByteDance Seed | Unclear | Public data only | 1 |
| Vina Classical docking program, not deep learning, and the field's de facto baseline | Scripps Research | Open | Public data only | 2 |
Small molecules (21)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| ADMET-AI ADMET prediction platform returning 41 endpoints per molecule against approved-drug distributions | Stanford University (computer science), Greenstone Biosciences | Open | Public data only | 1 |
| AQAffinity Open protein-ligand affinity model adding an affinity head to an open co-folding backbone | SandboxAQ | Open | Public data only | 1 |
| Beacon (Inductive Bio) Inductive Bio's closed ADMET/PK predictor trained on a pre-competitive pharma data consortium | Inductive Bio | Closed | Private or licensed data | 0 |
| Chemformer Pretrained molecule-string sequence model for reaction prediction and retrosynthesis, long used as a baseline | AstraZeneca (Molecular AI, Discovery Sciences) | Unclear | Private or licensed data | 1 |
| Chemprop v2 Open-source molecular property prediction library on directed message passing, the most common deep-learning baseline | MIT (chemical engineering and electrical engineering groups) | Open | Public data only | 0 |
| Enchant v1/v2 Closed multimodal transformer predicting molecular properties and clinically relevant endpoints from early data | Iambic Therapeutics | Closed | Own wet-lab training data | 0 |
| ether0 A 24B-parameter chemistry reasoning model whose answers are molecules | FutureHouse | Open | Public data only | 1 |
| KERMT Graph-transformer chemical pretrained model with multitask fine-tuning and acceleration for in-house ADMET prediction | Merck & Co. (modeling and informatics), NVIDIA (BioNeMo team) | Open | Private or licensed data | 0 |
| MiniMol Ten-million-parameter molecular foundation model densely pretrained on thousands of supervised tasks | Valence Labs (Recursion Pharmaceuticals) | Open | Public data only | 1 |
| MolE Molecular-graph foundation model adapting disentangled attention to graphs for property prediction | Recursion | Open | Public data only | 1 |
| MoLFormer-XL Chemical language model pretraining on billion-scale molecule strings, then fine-tuned for property prediction | IBM Research | Open | Public data only | 1 |
| Nesso-1 Open protein-ligand affinity model trading coarse-grained structure for inference speed | Valence Labs, Recursion Pharmaceuticals | Unclear | Public data only | 4 |
| ReaSyn v2 Synthesisability projection model that builds a synthesis route from purchasable building blocks | NVIDIA, KAIST | Open | Public data only | 1 |
| REINVENT 4 Open-source small-molecule generative design framework driven by reinforcement, transfer and curriculum learning | AstraZeneca Molecular AI (Discovery Sciences, R&D) | Open | Public data only | 0 |
| RetroDFM-R Reasoning-style retrosynthesis LLM trained with verifiable chemistry rewards, emitting precursors with a reasoning trace | Shanghai Jiao Tong University (X-LANCE Lab; School of Chemistry and Chemical Engineering), Suzhou Laboratory | Open | Public data only | 1 |
| Saturn Generative molecular design model aimed at sample-efficient search under a limited oracle budget | EPFL (Laboratory of Artificial Chemical Intelligence) | Open | Public data only | 0 |
| TamGen Target-aware molecule generator that conditions on a protein pocket and refines compounds | Microsoft Research AI for Science, Global Health Drug Discovery Institute (GHDDI) | Open | Public data + wet lab for validation only | 1 |
| TerraBind Protein-ligand structure and affinity foundation model trading diffusion for speed | Terray Therapeutics | Closed | Public data + wet lab for validation only | 2 |
| TxGemma A family of drug-discovery LLMs (three sizes, predict and chat variants) instruction-tuned from an open general base on a public therapeutics dataset library | Google DeepMind, Google Research | Partly open | Public data only | 1 |
| Uni-Mol2 3D-conformation molecular pretraining model scaled to the billion-parameter range to study scaling | DP Technology, Peking University, AI for Science Institute, Beijing | Open | Public data only | 1 |
| Uni-Mol3 Multi-molecular reaction foundation model covering products, retrosynthesis, conditions and yield | DP Technology, Peking University | Open | Private or licensed data | 1 |
Proteins (26)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| AlphaProteo Closed de novo protein binder design system; seven targets reported successful and an eighth failed | Google DeepMind, Francis Crick Institute | Closed | Public data + wet lab for validation only | 0 |
| AMPLIFY Small protein language models arguing data quality matters more than scale | Mila, Polytechnique Montréal, Amgen | Open | Public data only | 0 |
| BindCraft Open binder design pipeline that backpropagates through a co-folding network, with no extra training | EPFL Laboratory of Protein Design and Immunoengineering, MIT | Open | Public data + wet lab for validation only | 1 |
| BoltzGen Open general-purpose binder design model built on a co-folding architecture | MIT Jameel Clinic, Boltz PBC | Open | Public data + wet lab for validation only | 0 |
| Chai-3 Chai Discovery's closed next-generation molecular design model for antibodies and multispecifics | Chai Discovery | Closed | Private or licensed data | 0 |
| Claude protein-design study (general model + public tools) A general language model orchestrates public tools to design binders, tested by outside labs | Anthropic, Adaptyv Bio, Twist Bioscience | Closed | Public data + wet lab for validation only | 1 |
| DPLM-2 Multimodal discrete-diffusion protein language model generating matched sequence and structure | ByteDance Research, Nanjing University | Open | Public data only | 1 |
| Dyno Psi-1 / Psi-Phi Open-weight binder design model of about 200M parameters with a companion filter and agent suite | Dyno Therapeutics | Open | Public data + wet lab for validation only | 0 |
| ESM C Sequence-only protein language model positioned as a replacement for ESM-2 | Evolutionary Scale, Chan Zuckerberg Biohub | Open | Public data + wet lab for validation only | 1 |
| ESM3 Multimodal generative protein language model; sequence, structure and function prompt each other | Evolutionary Scale, Chan Zuckerberg Biohub | Open | Public data + wet lab for validation only | 4 |
| EVOLVEpro Few-shot active-learning wrapper for protein engineering, about a dozen variants per round | Brigham and Women's Hospital / Harvard Medical School, Mass General Brigham Gene and Cell Therapy Institute, Beth Israel Deaconess Medical Center, MIT | Open | Own wet-lab training data | 0 |
| La-Proteina All-atom protein generation with flow matching in a partially latent space, co-designable up to 800 residues | NVIDIA | Open | Public data only | 1 |
| Latent-X / Latent-X2 Closed platform models for binder, macrocycle and antibody design, testing only tens of designs per target | Latent Labs | Closed | Public data + wet lab for validation only | 0 |
| LucaOne Unified pretrained foundation model for DNA, RNA and protein sequences, about 1.8B parameters | Alibaba Cloud, Sun Yat-sen University | Open | Public data only | 0 |
| Mosaic Open multi-objective protein design library optimising binders through combined scorers | Escalante Bio | Open | Public data + wet lab for validation only | 1 |
| Pinal Natural-language-to-protein design: turns a function description into structure tokens, then a sequence | Westlake University | Open | Public data + wet lab for validation only | 0 |
| PoET-2 182M-parameter multimodal retrieval-augmented protein foundation model using family context | OpenProtein.AI | Partly open | Public data only | 1 |
| Profluent-E1 Retrieval-augmented protein encoder for mutation-effect scoring and representation | Profluent | Open | Public data only | 3 |
| ProGen3 Sparse mixture-of-experts autoregressive protein language model focused on scaling laws and alignment with laboratory data | Profluent | Partly open | Own wet-lab training data | 1 |
| ProSST Protein language model quantizing local structure into tokens with disentangled attention | Shanghai Jiao Tong University, Shanghai AI Laboratory | Open | Public data only | 2 |
| Proteina-Complexa All-atom binder design model combining generative pretraining with test-time compute | NVIDIA (GenAIR) | Open | Public data + wet lab for validation only | 1 |
| RFdiffusion2 / RFdiffusion3 Atomic-level diffusion design family: RFD2 scaffolds enzyme active sites, RFD3 designs all-atom interfaces | University of Washington Institute for Protein Design, Rosetta Commons | Open | Public data + wet lab for validation only | 1 |
| SaProt Structure-aware protein language model pairing structure tokens with amino acids | Westlake University, Westlake Representation Learning Lab | Open | Public data only | 3 |
| SeedProteo Diffusion-based all-atom de novo protein design model for binders | ByteDance Seed | Closed | Public data + wet lab for validation only | 0 |
| VenusREM Retrieval-augmented zero-shot mutation-effect model fusing sequence, structure and homologs | Shanghai Jiao Tong University, East China University of Science and Technology, Shanghai AI Laboratory | Open | Public data + wet lab for validation only | 1 |
| xTrimoPGLM A 100B unified protein language model from BioMap and partners for understanding and generation, with a folding branch | BioMap, Tsinghua University, MBZUAI | Partly open | Public data only | 4 |
Antibodies (7)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| Germinal Open-source de novo antibody and nanobody design pipeline that redesigns only the CDRs | Stanford University (chemical engineering), Arc Institute | Open | Public data + wet lab for validation only | 0 |
| HCAbLM Nona Biosciences language model trained only on fully human heavy-chain-only antibodies, aimed at developability | Nona Biosciences, HBM Holdings (Harbour BioMed) | Unclear | Own wet-lab training data | 0 |
| IgGM / IgGM-FM Generative model for antibodies and nanobodies covering structure prediction, de novo design, maturation and humanization | Tencent AI for Life Sciences Lab | Open | Public data + wet lab for validation only | 1 |
| JAM / JAM-2 Closed de novo antibody and VHH design models aimed at hard multi-pass membrane targets such as GPCRs | Nabla Bio | Closed | Public data + wet lab for validation only | 0 |
| Origin-1 Generative platform designing full-length antibodies de novo against zero-prior epitopes, combining diffusion generation, paired CDR sequence design and co-folding scoring | Absci | Partly open | Public data + wet lab for validation only | 0 |
| RFantibody De novo antibody design pipeline for specified epitopes: generate backbones, design sequences, filter by structure | University of Washington Institute for Protein Design | Open | Public data + wet lab for validation only | 0 |
| tFold Antibody structure and design system: MSA-free structure prediction plus epitope-specific de novo design | Tencent AI for Life Sciences Lab | Partly open | Public data + wet lab for validation only | 0 |
Genomics & RNA (9)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| AlphaGenome Sequence-to-function model reading 1 Mb of DNA for base-resolution regulatory tracks and variant effects | Google DeepMind | Open | Public data only | 3 |
| AlphaMissense Structure-augmented missense pathogenicity model with a large public variant catalogue | Google DeepMind | Open | Public data only | 2 |
| BioReason Multimodal model feeding a DNA foundation model into an LLM for pathway reasoning and variant effects | University of Toronto, Vector Institute, University Health Network | Open | Public data only | 1 |
| Borzoi Predicts RNA-seq coverage from DNA and scores variant effects on transcription, splicing, polyadenylation | Calico Life Sciences | Open | Public data only | 2 |
| BulkFormer Foundation model for human bulk transcriptomes at about 150 million parameters | Peking University (biomedical informatics), Jilin University, Wuhan Sports University | Open | Public data only | 0 |
| Evo 2 Open DNA language model with million-token context for zero-shot variant scoring and sequence generation | Arc Institute, NVIDIA, Stanford University, University of California, Berkeley, University of California, San Francisco | Open | Public data + wet lab for validation only | 2 |
| NTv3 Genome foundation model with 1 Mb context doing track prediction, gene annotation and sequence generation in one | InstaDeep (subsidiary of BioNTech) | Open | Public data + wet lab for validation only | 1 |
| popEVE Proteome-wide missense pathogenicity model with scores comparable across genes | Harvard Medical School (Blavatnik Institute), Centre for Genomic Regulation (CRG), Barcelona | Open | Public data only | 1 |
| RIBOSPAN 1.61B-parameter bidirectional RNA foundation model with a native 10,240-nt context and single-nucleotide tokens, able to take a full transcript at once | Shanghai Innovation Institute, CAS Center for Excellence in Molecular Cell Science, Shanghai Jiao Tong University | Open | Public data only | 1 |
Single-cell & perturbation (12)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| AIDO Cell Multi-scale, stateful virtual-cell world model, currently with two prototype cell types | GenBio AI | Closed | Not disclosed | 0 |
| C2S-Scale Single-cell model that turns ranked expression into "cell sentences" and fine-tunes general LLMs from 410M to 27B parameters | Yale University, Google Research, Google DeepMind | Open | Public data + wet lab for validation only | 1 |
| CellFM Human-only single-cell foundation model with 800 million parameters | Sun Yat-sen University (computer science), Chongqing University, Huawei, Singleron Biotechnologies | Open | Public data only | 1 |
| Lingshu-Cell Alibaba DAMO Academy generative cell world model using masked discrete diffusion over transcriptomes | Alibaba DAMO Academy | Closed | Public data only | 1 |
| scGPT / Geneformer First-generation single-cell foundation-model baselines for annotation, perturbation and integration | University of Toronto, University Health Network, Vector Institute, Gladstone Institutes, UCSF | Open | Public data + wet lab for validation only | 2 |
| scPRINT-2 Institut Pasteur's second-generation single-cell foundation model pretrained on 350M cells from 16 species | Institut Pasteur, CNRS | Open | Public data only | 0 |
| Stack Single-cell foundation model doing in-context learning from example cells in the prompt, without fine-tuning | Arc Institute | Open | Public data only | 1 |
| STATE Virtual cell model predicting transcriptional response to drug or genetic perturbation across cell lines | Arc Institute | Open | Public data only | 1 |
| Tahoe-x1 Single-cell foundation model: masked-pretraining encoder trained at scale including perturbation data | Tahoe Therapeutics | Open | Own wet-lab training data | 1 |
| TranscriptFormer CZI/Biohub cross-species generative single-cell model trained on about 110M cells from 12 species | Chan Zuckerberg Initiative, Chan Zuckerberg Biohub | Open | Public data only | 0 |
| X-Cell Virtual-cell model predicting expression change after genetic perturbation | Xaira Therapeutics | Closed | Own wet-lab training data | 1 |
| xTrimoSCPerturb / scFoundation BioMap's single-cell foundation model scFoundation and the VCC 2025-winning perturbation entry built on it | BioMap, Tsinghua University | Partly open | Public data only | 1 |
Molecule generation (1)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| GenMol NVIDIA's masked discrete-diffusion molecule generator covering several generation tasks in one model | NVIDIA | Open | Public data only | 1 |
Text, LLMs & agents (12)
| Model | Organization | Code | Training data | Benchmarks |
|---|---|---|---|---|
| Biomni General biomedical agent: an LLM plus a tool library that plans and runs bioinformatics analyses | Stanford University, Phylo | Open | Public data + wet lab for validation only | 3 |
| Co-Scientist Google DeepMind's Gemini-based multi-agent system that generates and ranks research hypotheses | Google DeepMind, Google Research | Closed | Public data + wet lab for validation only | 0 |
| Frontier general-purpose LLMs Category entry for general frontier LLMs (Claude, GPT, Gemini) tested on bio benchmarks | Anthropic, OpenAI, Google DeepMind | Closed | Not disclosed | 5 |
| GPT-Rosalind Life-sciences reasoning model offered through controlled access | OpenAI | Closed | Not disclosed | 1 |
| Intern-S Open multimodal scientific foundation series layering scientific corpora and specialised tokenisers on general LLMs, including a trillion-parameter mixture-of-experts version | Shanghai AI Laboratory | Open | Public data only | 0 |
| Kosmos / Edison Analysis Edison Scientific's "AI scientist" that loops literature search and data analysis into cited reports | Edison Scientific, FutureHouse | Closed | Public data only | 1 |
| MedGemma / Med-Gemini Google medical language-model family: closed Med-Gemini and open-weight MedGemma | Google Research, Google DeepMind | Partly open | Private or licensed data | 1 |
| Quine Microsoft Research's biology world model plus research harness, research use only | Microsoft Research | Closed | Public data + wet lab for validation only | 0 |
| Robin FutureHouse multi-agent discovery system linking literature review, candidate proposal and data analysis | FutureHouse | Open | Public data + wet lab for validation only | 0 |
| SciReasoner Multimodal science reasoning model encoding 3D structure as shared tokens | Shanghai AI Laboratory, The Chinese University of Hong Kong, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Stanford University, University of Oxford | Unclear | Public data only | 1 |
| TxAgent Therapeutic-reasoning agent calling 211 drug and clinical tools via ToolUniverse | Harvard Medical School | Open | Public data only | 0 |
| VERID Semantic digital twin of a lab workstation that loops agents between virtual dry runs, safety checks and real wet-lab experiments | Google Research | Unclear | Own wet-lab training data | 0 |