Hugging Face Releases Carbon-X, a 14B Open DNA Language Model Bridging Biology and Natural Language

Hugging Face has released Carbon-X, a 14-billion-parameter open DNA language model that, for the first time, jointly represents DNA sequences and natural language in a unified model. This multimodal biological approach allows the model to reason across both genomic data and human-readable text, enabling tasks like gene function annotation, biological question answering, and sequence-to-description mapping. For developers working at the intersection of AI and life sciences, Carbon-X offers an open-weight foundation model that can be fine-tuned for genomics pipelines without needing separate embedding systems for sequence and text data. The open release means researchers and engineers can run the model locally or via Hugging Face infrastructure, lowering the barrier to entry for computational biology applications significantly. This is a notable expansion of Hugging Face's model portfolio into scientific AI, following the broader trend of large labs releasing domain-specific foundation models.