Breaking

Hugging Face Launches Carbon-X, an Open-Source DNA Language Model for Genomic Research

October 23, 2026 at 02:19 EDT

Foundation Models
Research & Papers

Hugging Face has released Carbon-X, an open-source language model that can interpret DNA sequences and predict the effects of genetic mutations on gene activity and human health. The 14.8-billion-parameter model, made available on October 22, represents a significant step toward democratizing genomic analysis, allowing researchers to query genetic sequences in natural language rather than relying on specialized computational tools.

October 2026

Open DNA Language Model Scales Research Access

Carbon-X enables researchers to query genetic sequences in natural language, from local hardware, at a fraction of proprietary model costs.

Model Capacity

14.8B

Parameters trained on 302B tokens

Cost Efficiency

1,800×

Cheaper inference vs GPT-6.1 Sol

Benchmark Performance Comparison

Carbon-X

66.7%

GPT-6.1 Sol

67.5%

Variant-effect and sequence-understanding benchmarks

Key capability: Researchers query DNA sequences in natural language to predict mutation effects, identify regulatory regions, and classify genetic sequences—all locally on a single 80GB GPU.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.