Microsoft Research released Skala 1.1, a deep-learning density functional theory model that significantly improves accuracy in thermochemistry, reaction kinetics, noncovalent interactions, and molecular structure prediction while maintaining meta-GGA-level computational cost. Trained on 2.5 times more high-accuracy quantum-chemistry data from the Microsoft Research Accurate Chemistry Collection, Skala 1.1 surpasses leading global hybrid functionals on the GMTKN55 benchmark, achieving a weighted average error of 2.8 kcal/mol. Skala is now available in CP2K and being integrated into Psi4, FHI-aims, ORCA, and VASP, supported by a living benchmark to track performance and guide broader adoption in scientific and industrial computational workflows.
This update represents a notable development in the Ai sector. Organizations and founders tracking this space should evaluate potential strategic and technical implications on their operations.