Microsoft's Skala 1.1 Boosts Accuracy for Molecular Simulations

Microsoft Research has released Skala 1.1, an updated version of its deep-learning exchange-correlation functional for density functional theory (DFT) calculations. Trained on 2.5 times more data than its predecessor, Skala 1.1 shows improved accuracy in key areas including thermochemistry, reaction kinetics, and molecular structure prediction. The functional outperforms expensive hybrid functionals in 32 of 55 categories in the GMTKN55 benchmark while maintaining the computational cost of meta-GGA functionals. Microsoft has made Skala available in CP2K and is integrating it into Psi4, FHI-aims, ORCA, and VASP to expand access across the computational chemistry ecosystem. The company also introduced a living benchmark to track performance improvements across software and hardware platforms.