U.S. Energy Department launches Genesis open-weight models for science
The DOE has announced a new class of open-weight foundation models for scientific research, unveiling Genesis-Science-1 with Arcee and opening a contribution portal; first-round applications are due August 14, 2026.
A new class of open-weight models
The U.S. Department of Energy has announced the Genesis Open Models initiative, a new class of open-weight foundation models built to accelerate scientific discovery as part of the department's broader Genesis Mission. The models are intended for researchers, national laboratories, industry partners and the wider open-science community, and are meant to be transparent and extensible enough to be adapted to many scientific domains.
Genesis-Science-1, built with Arcee
The first model in the class is Genesis-Science-1, developed in partnership with Arcee, which is the first industry partner supporting the program. According to the initiative's website, releasing models with open weights is meant to rally the scientific and AI communities around shared infrastructure for science.
Where the models will be used
The DOE lists materials discovery, energy systems, earth systems modelling, fusion, biology and high-energy physics among the target areas, and grounds the effort in open science, reproducibility and responsible AI, with the stated goal of lowering the barriers to advanced AI capabilities for the public good.
Open portal, contributions due August 14
A DOE-hosted contribution portal has opened for parties that want to take part, with first-round applications due August 14, 2026. The department is collecting three kinds of contributions: open-weight models, both as base models for downstream fine-tuning and for immediate deployment, provided they have transparent provenance; pretraining contributions from organisations and researchers holding high-quality, domain-specific scientific data, such as curated datasets, benchmarks or specialised corpora; and fine-tuning efforts that shape versions of Genesis-Science-1 for particular fields, from lab assistants to simulation surrogates and scientific copilots. The site says participants can also co-develop evaluations and benchmarks, and that their engagement will help shape the capabilities, governance and impact of the Genesis models as a shared public asset.
SiTech — AI-powered web development
We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.