Hugging Face has launched Carbon-X, a 14.8 billion parameter open-source DNA language model. This model integrates DNA sequences and natural language, allowing it to process raw DNA data and respond in English. Carbon-X can predict the pathogenicity of mutations, identify regulatory regions, classify organisms, and predict the next DNA base. It is built upon the Qwen3-14B architecture and features a hybrid tokenizer that combines standard text tokens with specialized DNA tokens. AI
IMPACT Enables new research avenues by integrating DNA analysis with natural language processing capabilities.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Apache Software License 2.0
- BioReason
- Carbon X-ray absorption spectra of fluoroethenes and acetone: a study at the coupled cluster, density functional, and static-exchange levels of theory.
- deoxyribonucleic acid
- Gemini 4 Argon
- GPT-6.1 Sol
- Hugging Face
- Opus 5.5
- Qwen3 14B
- University of Naples Federico II
- Zhongguancun Academy
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
