88Z AI, 4 August — ModelBest and OpenBMB have open-sourced ForgeStencil, the world's first AI-driven automated Stencil optimization system. By leveraging the collaboration of two agents, Kernel Agent and App Agent, the system achieves a fully automated workflow from operator research to application deployment without human intervention. ForgeStencil demonstrates high efficiency by completing end-to-end optimization for over 100 real-world industrial and scientific software programs per week, covering eight major i
At a glance
- The world's first AI-driven automated Stencil optimization system uses dual agents to achieve zero human intervention.
- Optimizes over 100 real software programs weekly across eight industrial sectors including oil, electromagnetics, and healthcare.
The update
ModelBest and OpenBMB have jointly announced the open-source release of ForgeStencil, marking the debut of the world's first AI-driven automated Stencil optimization system. The core innovation of this system lies in its dual-agent architecture, featuring Kernel Agent and App Agent. These two agents work in close coordination to revolutionize traditional development models, achieving full automation from low-level operator research to high-level application deployment without any human intervention. In terms of practical performance, ForgeStencil demonstrates robust capabilities for industrial implementation. The system is capable of completing end-to-end optimization tasks for more than 100 real-world industrial and scientific software programs every week. Its application scope is extensive, deeply covering eight key industrial sectors such as oil and gas exploration, electromagnetic simulation, and medical imaging, thereby providing efficient automated solutions for high-performance computing scenarios.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 面壁智能Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.