Hugging Face Unveils Open-source AI for Natural-language DNA Queries

WHO:

Techmeme - Hugging Face released the Carbon-X model.

Wired Science - The article identifies Hugging Face’s new AI biology model as an open-source model.

WHAT:

Techmeme - Carbon-X is an open-source, multimodal AI model with 14 billion parameters, trained on 300 billion tokens.

Techmeme - The model lets users query DNA sequences using natural language.

Wired Science - The model “speaks the language of DNA” and can help researchers understand how changes affect gene activity and health.

WHY:

Wired Science - The model is intended to help researchers understand the effects of changes to DNA on gene activity and health.

HOW:

Techmeme - Carbon-X uses multimodal AI to let users ask natural-language questions about DNA sequences.

Wired Science - The model enables researchers to interact with DNA using language.

(UPDATED October 22, 2026 12:03 UTC)

Hugging Face has released Carbon-X, an open-source AI model designed to help researchers explore DNA. The multimodal model has 14 billion parameters and was trained on 300 billion tokens. It allows users to ask natural-language questions about DNA sequences, making it possible to interact with genetic data using everyday language. The model is described as speaking the language of DNA. Researchers can use it to investigate how changes in a DNA sequence may affect gene activity and health. By bringing language-based queries to DNA analysis, Carbon-X aims to make it easier to examine the possible consequences of genetic changes. Its open-source release also makes the model available for others to use and build on.