# Meta's Open Vision Models Run Inside a Department of Energy Beamline Pipeline and an ARPA-H Wheelchair Robot

Berkeley Lab says SAM 3 and DINOv3 on 300 A100 GPUs cut beamline image analysis from months to about 15 minutes, and both projects fine-tune the models on their own infrastructure.

- Published: 2026-11-09T06:17:39.329Z
- Canonical: https://polylog.news/ai/2026-11-09/meta-s-open-vision-models-run-inside-a-department-of-energy
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Meta AI](https://ai.meta.com/blog/genesis-mission-lawrence-berkeley-national-laboratory-segment-anything-dino/), [Meta AI](https://ai.meta.com/blog/assistive-robotics-university-of-pittsburgh-sam-dino/)

SYNAPS-I is one of the flagship projects of the Department of Energy's (DOE) Genesis Mission. Berkeley Lab leads it with Argonne, Brookhaven, Oak Ridge and SLAC, and its segmentation pipeline is built on two Meta models: Segment Anything Mo…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-11-09/meta-s-open-vision-models-run-inside-a-department-of-energy (subscription information: https://polylog.news/pricing).