Announcing Genesis-Science-1, an Open-Weight Model for Scientific Research

The U.S. Department of Energy and Arcee AI today announced Genesis-Science-1, or GS1, an American open-weight AI model and governed research system designed to complete scientific computing workflows while preserving a reproducible record of its work.

GS1 is a new program under DOE’s Genesis Mission. Arcee AI will lead model development. DOE scientists and engineers at participating national laboratories will supply reviewed scientific materials, define representative research tasks, design evaluations, and validate the system’s results.

Arcee will secure the compute, curate the training data, train the model through pretraining and post-training, build the governed execution environment, create training workbenches from approved DOE materials, and conduct evaluation and release.

DOE and participating national laboratories will ground the system in scientific practice. Contributions may include experimental and observational data from user facilities, simulation outputs and run logs from supercomputing campaigns, materials and chemistry collections, research software, and the tools scientists use to conduct and review their work. Materials will enter the program only after completing the Department’s release-review process.

DOE also opened the Genesis-Science-1 contribution portal today. The portal, hosted by Argonne National Laboratory will open later this week and invites universities, national laboratories, companies, scientific nonprofits, and research organizations to contribute data, research environments, evaluations, and technical expertise. Applications for the first contribution window close August 6, 2026.

"A country cannot lead in AI if everything it leads in is closed," said Mark McQuade, co-founder and CEO of Arcee AI. "With Genesis-Science-1, we're holding American open weights to a demanding standard: useful scientific work under real operating constraints, judged by the people who do it."

Scientific computing rarely begins with a clean prompt and a single correct answer. A researcher may inherit an aging Fortran codebase, a partially completed simulation campaign, conflicting run logs, and several reasonable options for what to try next. Progress depends on a sequence of technical judgments about which result to trust, which test to run, when to restart, and whether the evidence supports a conclusion.

GS1 will train in scientific workbenches that reproduce these working conditions. Initial areas include high-performance-computing code modernization, experimental analysis, simulation campaigns, materials science, and energy systems.

Each workbench will contain the code, data, tools, documentation, logs, partial results, and failure states needed to reconstruct a research workflow. Training environments may include Python, Fortran, C and C++, MPI and OpenMP, CUDA and HIP, command-line tools, notebooks, simulation packages, and computing schedulers.

The model will operate through a governed execution system. Approved tools will run in sandboxed, staged environments. The system will maintain task state, checkpoint progress, manage retries and recovery, and record the prompts, tool calls, code changes, datasets, intermediate artifacts, and conclusions associated with each run.

Human review will remain part of the process. People will approve decisions involving safety, security, publication, and resource use. GS1 will not receive blanket access to DOE systems.

A successful run must carry a workflow from plan through report, revise the approach when the evidence changes, recover from tool failures, and leave a record that another researcher can inspect. Scientific experts will judge whether the result is sound and whether the supporting evidence is complete enough to reproduce.

DOE and its laboratories may need to run a model inside their own infrastructure, preserve a specific version for years, adapt it to a specialized field, and operate without a permanent dependency on an external API.

Open weights support those requirements by allowing an institution to hold and operate the model directly. They do not eliminate the need for permissions, evaluations, sandboxing, audit logs, deployment controls, or disciplined operations. The institution responsible for the work must still design and enforce those controls.

GS1 is intended to increase the supply of American open-weight models built for those operating conditions. Its release will include the model weights, a technical report, and public workbench and demonstration artifacts.

Arcee has built and released open-weight models end to end on compressed timelines, with experience spanning data, architecture, pre-training, post-training, evaluation, and deployment. Over six months, the company scaled the Trinity program through increasingly large training runs, culminating in Trinity Large, a 400-billion-parameter sparse mixture-of-experts model. That operating experience allows Arcee to work closely with DOE scientists and contributors throughout development and deliver Genesis-Science-1 on an accelerated schedule.

The program is seeking scientific and technical materials for several stages of model development.

Foundation-stage contributions may include scientific text, code, documentation, and structured technical collections suitable for pretraining, midtraining, or context extension.

Post-training contributions may include expert demonstrations, annotated task examples, research software, datasets, complete workflow environments, reinforcement-learning tasks, held-out evaluations, scoring rubrics, tests, and verifiers.

The program also needs scientists and engineers who can review tasks, model behavior, and completed artifacts. Infrastructure providers may submit rolling inquiries about additional training or evaluation capacity.

The 2026 program has two scheduled contribution tracks. Organizations may apply to both when the proposed contributions are distinct.

Applicants will identify the relevant track, describe the scientific work and its importance, explain what material is ready, name the experts available to support it, and state the proposed terms of use.

The public application form collects descriptions and metadata only. No scientific material is transferred during the application process, and each contributor specifies how its material may be handled.

Applications will be reviewed through five gates: scientific fit, rights and handling, expert and evaluation readiness, technical integration, and final program selection.

Selected contributors will work with the researchers building and evaluating GS1. They will receive early evaluation access as the system develops and will be credited in the technical report and release materials.

The program is particularly interested in workflows that are poorly represented by standard AI evaluations and require direct collaboration with the researchers who perform the work.

The DOE contribution portal opens this week. Check back here for the link.

Questions may be directed to: genesis-science-1@anl.gov or genesismission@arcee.ai

Launched by executive order in November 2025, the Genesis Mission is a Department of Energy-led national initiative to accelerate scientific discovery through artificial intelligence. Its stated goal is to double the productivity and impact of American science and engineering within a decade.

The mission connects DOE’s 17 national laboratories, leadership-class computing facilities, scientific data resources, and partners across industry and academia. DOE Under Secretary for Science Darío Gil directs the Genesis Mission. More information is available at energy.gov.

Arcee AI is a U.S. open-model lab that develops open-weight foundation models end to end, including data preparation, architecture, pretraining, post-training, evaluation, and deployment.

Its Trinity family ranges from compact models designed for local use to Trinity Large, a 400-billion-parameter sparse mixture-of-experts model. Arcee builds models that developers, enterprises, and public institutions can run and adapt on infrastructure they control. More information is available at arcee.ai.