Carbon-X is an open 14.8B model trained jointly on DNA and text [R]
Summary
Carbon-X is an open 14.8B decoder-only model that jointly processes raw DNA and text via a shared 6-mer token stream, achieving 66.7% average on five DNA benchmarks — nearly matching GPT-6.1 Sol — with Apache 2.0 weights that fit on a single 80GB GPU.
Similar Articles
Carbon-X: the best open DNA language model runs locally with llama.cpp
HuggingFace Bio 发布 Carbon-X，一个 14B 的 DNA+自然语言联合模型，在五项 DNA 基准上超越所有开源模型并匹配 GPT-6.1 Sol，且成本低约 1000 倍，同时提供 GGUF 版本支持 llama.cpp 本地运行。
Carbon: Decoding the Language of Life
Hugging Face released Carbon, a family of open DNA foundation models that matches state-of-the-art performance of Evo2-7B while being 275x faster, using 6-mer tokenization, factorized loss, and curated genomic data.
@aisolopreneur: Hugging Face just released Carbon-X, an open model you can ask questions about DNA in plain English, and on its own tes…
Hugging Face released Carbon-X, an open DNA language model that answers questions about gene sequences in plain English, scoring 66.7% on DNA benchmarks near GPT-6.1 Sol at roughly 1/1,800 the cost per problem. It runs on a single 80GB GPU, potentially making large-scale variant classification dramatically cheaper for hospitals and researchers, though it is not yet clinically validated.
@lvwerra: We are releasing Carbon: a crazy fast DNA model Carbon is 275x faster than the next best model. So fast you can process…
HuggingFace releases Carbon, a DNA model that is 275x faster than the previous state-of-the-art (Evo2), enabling processing of the entire human genome on a single GPU in under two days. The model uses a unique tokenizer that splits sequences into 6-base chunks while maintaining single-base resolution, and comes with an interactive demo.
HuggingFaceBio/Carbon-X (DNA understanding and generation, Qwen 3 finetune)
HuggingFaceBio 发布开源 14.8B DNA 理解与生成模型 Carbon-X，基于 Qwen3-14B 微调，融合 DNA 序列与自然语言训练，在五项 DNA 基准上平均得分 66.7%，接近前沿闭源模型但推理成本约为其 1/1800。