Company Updates

Los Alamos' Venado-GPT processes 3B+ tokens weekly as OpenAI partnership hits one-year mark

Share
Los Alamos' Venado-GPT processes 3B+ tokens weekly as OpenAI partnership hits one-year mark

Los Alamos National Laboratory and OpenAI have marked one year of partnership, with Venado-GPT's weekly token throughput growing 20-fold to over 3 billion. The collaboration has significantly enhanced research capabilities across three national laboratories.

TL;DR

  • Venado-GPT's usage has surged, processing over 3 billion tokens weekly and tripling user accounts.
  • The partnership enables advanced research in scientific and national security domains.
  • Future enhancements include improved file attachment reading and API dependability.

What happened

One year into their partnership, Los Alamos National Laboratory and OpenAI report substantial growth in the use of Venado-GPT, a service that hosts OpenAI's experimental SI models on Los Alamos' Venado supercomputer. Venado-GPT is accessible to staff at Los Alamos, Lawrence Livermore, and Sandia national laboratories, collectively known as the tri-laboratories.

From mid-February to late August 2026, Venado-GPT's weekly token throughput increased 20-fold to over 3 billion tokens. The number of user accounts tripled, and weekly successful requests rose from approximately 1,200 to more than 41,200. The models, named 5.4 OpenAI, 5.4 OpenAI mini, and o3 OpenAI, run on Venado's local GPUs, linked by the HPE Slingshot interconnect.

Why it matters

The partnership has enabled new research capabilities, applying AI models to solve pressing scientific and national security problems. Researchers have used Venado-GPT to streamline workflows, advance mission-relevant work, and develop automated agent-driven translation of simulation models.

For AI/ML developers and startup founders, this collaboration highlights the potential of integrating advanced AI models with high-performance computing to tackle complex challenges. Tech investors may see this as a testament to the growing importance of AI in scientific research and national security.

Key facts

  • Venado-GPT's weekly token throughput grew 20-fold to over 3 billion tokens in six months.
  • The number of tri-laboratory user accounts tripled.
  • Weekly successful requests increased from approximately 1,200 to over 41,200.
  • Venado-GPT models are used in tandem with the Universal Research and Scientific Agent (URSA) for mission-relevant work.
  • Lawrence Livermore researchers use 5.4 OpenAI for automated agent-driven translation of simulation models.
  • Future enhancements include improving file attachment reading and API dependability.
  • The partnership is preparing for the arrival of the upcoming Mission and Vision platforms.

Context

This collaboration between Los Alamos National Laboratory and OpenAI exemplifies the increasing integration of AI in scientific research and national security. It underscores the importance of high-performance computing and advanced AI models in addressing complex challenges.

For the AI industry, this partnership serves as a model for how AI can be leveraged to enhance research capabilities and drive innovation. It also highlights the growing role of AI in the public sector and national security.

Topics

Join the discussion

Have a take on this story? Weigh in with our community on Facebook.

💬 Discuss on Facebook →