The open platform where genomic AI researchers collaborate on models, datasets, and tools for variant interpretation, gene-disease associations, and precision medicine.
OpenGenomeLLM-70BGenomicsclassify_variant(vcf)VCF→ReportClinVar: Pathogenic ✓ValidatedFlagship open-source genomic LLM for variant interpretation, gene-disease associations, and clinical report generation from VCF files. 78.9% on GeneTuring.
State-of-the-art biomedical LLM — 91.2% on MedQA. Trained on 42M+ PubMed abstracts with DPO alignment on physician-curated preference data.
Compact biomedical LLM for on-premise and edge clinical systems. GGUF and AWQ quantized variants available. Runs on a single RTX 4090.
Clinical decision support model fine-tuned on de-identified EHR data and clinical pathways. Differential diagnosis generation and treatment recommendations.
Specialized for genomic sequence analysis and variant interpretation. Trained on NCBI, Ensembl, ClinVar, and OMIM. 87% on variant classification benchmark.
ADMET property prediction, SMILES-to-property conversion, and compound-target interaction modeling. Trained on 10M+ compounds from ChEMBL and PubChem.
Download and run any OpenGenomeLLM model locally using HuggingFace Transformers. All models are Apache 2.0 licensed for research and commercial use with no restrictions.
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
# Load OpenGenomeLLM-70B
tokenizer = AutoTokenizer.from_pretrained(
"deepcog-ai/OpenGenomeLLM-70B"
)
model = AutoModelForCausalLM.from_pretrained(
"deepcog-ai/OpenGenomeLLM-70B",
torch_dtype=torch.bfloat16,
device_map="auto"
)
# Interpret a genomic variant
query = "Interpret BRCA1 c.5266dupC for clinical significance"
inputs = tokenizer(query, return_tensors="pt")
output = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(output[0]))Find the right model for your specific research or clinical workflow.
Evaluated on GeneTuring, ClinVar VUS classification, and Gene-Disease Association tasks against leading models.
Compare All Models →Create a free account to download models, contribute datasets, and collaborate with the global genomic AI community.