{"id":1184167,"date":"2026-08-20T09:00:00","date_gmt":"2026-08-20T16:00:00","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?p=1184167"},"modified":"2026-08-20T07:41:28","modified_gmt":"2026-08-20T14:41:28","slug":"broadening-access-to-skala-creates-a-faster-path-to-predictive-dft","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/research\/blog\/broadening-access-to-skala-creates-a-faster-path-to-predictive-dft\/","title":{"rendered":"Broadening access to Skala creates a faster path to predictive DFT\u00a0"},"content":{"rendered":"\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1400\" height=\"788\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1.jpg\" alt=\"Schematic of the Skala architecture, showing how meta-GGA electronic features are transformed through point-wise processing and non-local atomic interactions to predict density functional theory energies.\" class=\"wp-image-1184230\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1.jpg 1400w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-1280x720.jpg 1280w\" sizes=\"auto, (max-width: 1400px) 100vw, 1400px\" \/><\/figure>\n\n\n\n<div style=\"padding-bottom:0;padding-top:0\" class=\"wp-block-msr-immersive-section alignfull row\">\n\t\n\t<div class=\"container\">\n\t\t<div class=\"wp-block-msr-immersive-section__inner wp-block-msr-immersive-section__inner--narrow\">\n\t\t\t<div class=\"wp-block-columns mb-10 pb-1 pr-1 is-layout-flex wp-container-core-columns-is-layout-8f761849 wp-block-columns-is-layout-flex\" style=\"box-shadow:var(--wp--preset--shadow--outlined)\">\n<div class=\"wp-block-column is-layout-flow wp-block-column-is-layout-flow\">\n<h2 id=\"at-a-glance\" class=\"wp-block-heading h3\">At a glance<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Skala 1.1 demonstrates the continuously improving nature of Microsoft Research&#8217;s deep-learning DFT approach: trained on 2.5\u00d7 more data than its predecessor, it delivers substantially higher accuracy across key molecular simulation challenges, including thermochemistry, reaction kinetics, and molecular structure prediction.<\/li>\n\n\n\n<li>Skala is now available in <strong>CP2K<\/strong> and is being integrated into <strong>Psi4<\/strong>, <strong>FHI-aims<\/strong>, <strong>ORCA<\/strong> and <strong>VASP<\/strong>, bringing next-generation DFT accuracy closer to the communities that rely on these codes every day.<\/li>\n\n\n\n<li>Microsoft Research is also introducing a living benchmark that will track the computational performance of successive, increasingly optimized Skala releases to help the community measure and accelerate progress toward ever greater accuracy and efficiency.<\/li>\n\n\n\n<li>Together, these developments mark another milestone toward a future in which computational chemistry simulations are both predictive and integrated in all relevant scientific and industrial workflows.<\/li>\n<\/ul>\n<\/div>\n<\/div>\t\t<\/div>\n\t<\/div>\n\n\t<\/div>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Bringing density functional theory (DFT) to predictive accuracy is a journey, not a single breakthrough.<\/strong> Since <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/blog\/breaking-bonds-breaking-ground-advancing-the-accuracy-of-computational-chemistry-with-deep-learning\/\" target=\"_blank\" rel=\"noreferrer noopener\">introducing Skala<\/a>, our deep-learning exchange-correlation functional, we have continued to advance along two complementary fronts: improving accuracy and expanding accessibility across the computational chemistry ecosystem.<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1785\" height=\"2560\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-scaled.png\" alt=\"Fig. 1: Table of errors for Skala-1.1 and competing density functionals on the 55 subsets of GMTKN55. Skala-1.1 delivers the lowest error on 32 subsets, indicating broad and consistent accuracy across a wide range of chemical properties and reaction types.\" class=\"wp-image-1184173\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-scaled.png 1785w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-209x300.png 209w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-714x1024.png 714w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-768x1101.png 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-1071x1536.png 1071w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-1428x2048.png 1428w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig1-blog_SKALA1.1-126x180.png 126w\" sizes=\"auto, (max-width: 1785px) 100vw, 1785px\" \/><figcaption class=\"wp-element-caption\">Figure 1: Accuracy of Skala-1.1 for thermochemistry, kinetics, and non-covalent interactions. At the computational cost of a meta-GGA functional, Skala 1.1 outperforms the best, most expensive global hybrid functionals, ranking first (earning gold medals) in 32 of the 55 categories of the widely used GMTKN55 benchmark, which spans a broad range of chemical problems.<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">On the accuracy front, the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2506.14665\" target=\"_blank\" rel=\"noopener noreferrer\">release of <strong>Skala-1.1<\/strong><span class=\"sr-only\"> (opens in new tab)<\/span><\/a> provides the first demonstration of the continuous-improvement paradigm underlying Skala. Trained on 2.5x more data than the first public version of Skala, the updated&nbsp; model delivers substantially improved performance across key challenges in molecular simulation, including main-group thermochemistry, reaction kinetics, and molecular structure prediction.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">But accuracy alone is not enough. DFT is the computational engine behind a vast range of scientific and industrial workflows, spanning chemistry, materials science, catalysis, energy technologies, and drug discovery. To have real-world impact, advanced functionals must be accessible where scientists already perform their calculations. That is why we are also expanding the Skala ecosystem through collaborations with leading electronic-structure software developers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Today, we are announcing that Skala is available in <strong>CP2K<\/strong> and is being integrated into <strong>Psi4<\/strong>, <strong>FHI-aims<\/strong>, <strong>ORCA<\/strong> and <strong>VASP<\/strong>,&nbsp; bringing next-generation DFT accuracy closer to the communities that rely on these codes every day. Alongside these integration efforts, we are introducing a living benchmark that tracks the computational performance of successive, increasingly optimized Skala releases. By providing a transparent and continuously updated reference for implementations across software packages and hardware platforms, this resource will help the community measure and accelerate progress toward ever greater accuracy and efficiency.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Together, these developments mark another milestone toward a future in which computational chemistry simulations are both predictive and accessible across a broader range of relevant scientific and industrial workflows.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Want to learn more about Skala and why DFT plays such an important role in in-silico discovery? Read also <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/aka.ms\/skaladft\/blog\" type=\"link\" id=\"https:\/\/aka.ms\/skaladft\/blog\" target=\"_blank\" rel=\"noopener noreferrer\">our first blog post<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>.<\/p>\n\n\n\n<h2 id=\"skala-as-a-continuously-improving-functional\" class=\"wp-block-heading\">Skala as a continuously improving functional<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Unlike the traditional &#8220;functional zoo&#8221;, where new functionals accumulate without replacing older ones, Skala follows a different philosophy: each release is designed to supersede the previous one. As new data, model architectures, and training strategies become available, the model improves while maintaining the same practical computational cost.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Skala-1.1<\/strong> is the latest demonstration of this approach. It achieves a weighted average error of <strong>2.8 kcal\/mol on GMTKN55<\/strong>, a widely used benchmark suite comprising 55 categories of chemistry, including thermochemistry, reaction barriers, and noncovalent interactions. This level of accuracy surpasses today&#8217;s leading global (range-separated) hybrid functionals while retaining the efficiency of a semi-local functional. Beyond energies, Skala-1.1 also provides highly accurate electron densities, dipole moments, and molecular geometries.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">These advances were enabled by major expansions of the <strong><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.nature.com\/articles\/s41597-026-07200-8\" type=\"link\" id=\"https:\/\/www.nature.com\/articles\/s41597-026-07200-8\" target=\"_blank\" rel=\"noopener noreferrer\">Microsoft Research Accurate Chemistry Collection<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> (MSR-ACC)<\/strong>, our large-scale collection of high-accuracy quantum-chemistry reference data generated with expensive wavefunction methods. For Skala-1.1, we added new categories, including electron affinities and noncovalent clusters, increasing both the size and, crucially, the diversity of the training data. This data-driven approach allows Skala to improve systematically with each generation, moving us closer to a truly scalable and predictive DFT framework.<\/p>\n\n\n\n<h2 id=\"available-where-scientists-work\" class=\"wp-block-heading\">Available where scientists work<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">To fully realize the potential of Skala\u2019s continuously evolving approach to DFT, we need dedicated infrastructure that allows new releases to be rapidly and seamlessly integrated into the major software packages used by scientists in industry and academia. In turn, this will establish the fast feedback loop essential for accelerating Skala\u2019s ongoing development.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We first made Skala available through our <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/microsoft\/skala\" type=\"link\" id=\"https:\/\/github.com\/microsoft\/skala\" target=\"_blank\" rel=\"noopener noreferrer\">open-source community release<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, built on (GPU4)<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/pyscf.org\/\" type=\"link\" id=\"https:\/\/pyscf.org\/\" target=\"_blank\" rel=\"noopener noreferrer\">PySCF<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> and <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/aka.ms\/SkalaDFT\/ASE\" type=\"link\" id=\"https:\/\/aka.ms\/SkalaDFT\/ASE\" target=\"_blank\" rel=\"noopener noreferrer\">integrated with ASE<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>. This enables researchers to evaluate and apply Skala with minimal effort while benefiting from highly optimized CPU and GPU performance.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">But no single software package can meet the needs of every application or research community. Computational chemistry and materials science rely on a rich ecosystem of electronic-structure codes, each shaped over decades to tackle specific scientific and industrial challenges. Bringing Skala to this broader ecosystem has therefore been a major focus of the past year. We are fortunate to build on the remarkable foundations created by the DFT community and grateful to the many researchers and developers who are helping to make Skala available within the software platforms that scientists use every day.<\/p>\n\n\n\n\t<div class=\"border-bottom border-top border-gray-300 mt-5 mb-5 msr-promo text-center text-md-left alignwide\" data-bi-aN=\"promo\" data-bi-id=\"999693\">\n\t\t\n\n\t\t<p class=\"msr-promo__label text-gray-800 text-center text-uppercase\">\n\t\t<span class=\"px-4 bg-white display-inline-block font-weight-semibold small\">Spotlight: Event Series<\/span>\n\t<\/p>\n\t\n\t<div class=\"row pt-3 pb-4 align-items-center\">\n\t\t\t\t\t\t<div class=\"msr-promo__media col-12 col-md-5\">\n\t\t\t\t<a class=\"bg-gray-300 display-block\" href=\"https:\/\/www.microsoft.com\/en-us\/research\/event\/microsoft-research-forum\/past-episodes\/?OCID=msr_researchforum_MCR_Blog_Promo\" aria-label=\"Microsoft Research Forum\" data-bi-cn=\"Microsoft Research Forum\" target=\"_blank\">\n\t\t\t\t\t<img decoding=\"async\" class=\"w-100 display-block\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2025\/05\/Research-Forum-hero_1400x788.jpg\" alt=\"Research Forum | abstract background with colorful hexagons\" \/>\n\t\t\t\t<\/a>\n\t\t\t<\/div>\n\t\t\t\n\t\t\t<div class=\"msr-promo__content p-3 px-5 col-12 col-md\">\n\n\t\t\t\t\t\t\t\t\t<h2 class=\"h4\">Microsoft Research Forum<\/h2>\n\t\t\t\t\n\t\t\t\t\t\t\t\t<p id=\"microsoft-research-forum\" class=\"large\">Join us for a continuous exchange of ideas about research in the era of general AI. Watch the latest episodes on demand.<\/p>\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"wp-block-buttons justify-content-center justify-content-md-start\">\n\t\t\t\t\t<div class=\"wp-block-button\">\n\t\t\t\t\t\t<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/event\/microsoft-research-forum\/past-episodes\/?OCID=msr_researchforum_MCR_Blog_Promo\" aria-describedby=\"microsoft-research-forum\" class=\"btn btn-brand glyph-append glyph-append-chevron-right\" data-bi-cn=\"Microsoft Research Forum\" target=\"_blank\">\n\t\t\t\t\t\t\tWatch on-demand\t\t\t\t\t\t<\/a>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><!--\/.msr-promo__content-->\n\t<\/div><!--\/.msr-promo__inner-wrap-->\n\t<\/div><!--\/.msr-promo-->\n\t\n\n\n<h2 id=\"from-community-release-to-native-integrations\" class=\"wp-block-heading\">From community release to native integrations<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">In collaboration with the team of Prof. Thomas D. K\u00fchne at the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/casus.science\" type=\"link\" id=\"https:\/\/casus.science\" target=\"_blank\" rel=\"noopener noreferrer\">Center for Advanced Systems Understanding (CASUS)<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, Skala has been successfully integrated into the open-source <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.cp2k.org\/\" type=\"link\" id=\"https:\/\/www.cp2k.org\/\" target=\"_blank\" rel=\"noopener noreferrer\">CP2K<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> package. With more than 25 years of development, CP2K is a powerhouse for DFT simulations, particularly for large-scale systems and long-timescale molecular dynamics, while also providing a rich portfolio of high-accuracy electronic-structure methods. Skala expands the frontiers of what is possible within CP2K, delivering a step change in DFT accuracy while preserving the computational efficiency needed for simulations at scale. We are excited to see how CP2K&#8217;s scale and versatility, combined with Skala&#8217;s continuously improving accuracy, will enable new scientific applications and discoveries in the years ahead.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">There is more to come. Together with its vibrant developer\u2019s community , we are actively integrating Skala into the open-source <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/psicode.org\/\" type=\"link\" id=\"https:\/\/psicode.org\/\" target=\"_blank\" rel=\"noopener noreferrer\">Psi4<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> package, an essential platform for molecular electronic-structure research. Combined with the PySCF-based Skala Community Edition, this will make Skala available in three widely used open-source quantum chemistry packages.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Beyond open-source software, we are working closely with leading developers behind <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.fhi-aims.org\/\" target=\"_blank\" rel=\"noopener noreferrer\">FHI-aims<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.faccts.de\/orca\/\" target=\"_blank\" rel=\"noopener noreferrer\">ORCA<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, and <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/vasp.at\/\" target=\"_blank\" rel=\"noopener noreferrer\">VASP<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, with the goal of making Skala broadly accessible across the major software platforms used in computational chemistry and materials science.<\/p>\n\n\n\n<h2 id=\"validating-accuracy-across-implementations-cp2k-as-case-study\" class=\"wp-block-heading\">Validating accuracy across implementations: CP2K as case study<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Thorough testing is essential for any new implementation. We want to ensure that Skala delivers consistent accuracy across different codes and computational settings. Together with the CP2K team, we developed a comprehensive suite of integration tests to verify that Skala produces numerically correct and reliable results. We are particularly grateful to the CASUS team, whose deep expertise in the numerical verification of computational methods was instrumental in designing and validating this testing framework.<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full is-resized\"><img loading=\"lazy\" decoding=\"async\" width=\"2007\" height=\"2007\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1.png\" alt=\"Fig. 2: Plot comparing signed errors for Skala-1.1 using the CP2K and PySCF implementations on a representative subset of GMTKN55. The two implementations show nearly identical results, agreeing within 0.04 kcal\/mol and confirming that the CP2K integration reproduces the accuracy of the community release.\" class=\"wp-image-1184191\" style=\"width:556px;height:auto\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1.png 2007w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-300x300.png 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-1024x1024.png 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-150x150.png 150w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-768x768.png 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-1536x1536.png 1536w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-180x180.png 180w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig2_SKALA1.1-360x360.png 360w\" sizes=\"auto, (max-width: 2007px) 100vw, 2007px\" \/><figcaption class=\"wp-element-caption\">Figure 2: Signed errors relative to high-accuracy reference values for a representative subset of GMTKN55, comparing the CP2K and PySCF implementations of Skala-1.1 using as closely matched numerical settings as possible. The two implementations agree to within 0.1 kcal\/mol MAD across the entire subset, with a single outlier corresponding to a particularly challenging radical system.&nbsp;<\/figcaption><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">A detailed discussion of the implementation, validation strategy, and testing infrastructure for Skala in CP2K can be found in our joint paper with the CASUS team: \u201c<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/publication\/molecular-implementation-of-the-machine-learned-skalaexchange-correlation-functional-in-cp2k-through-gauxc\/\">Molecular Implementation of the Machine-Learned Skala Exchange-Correlation Functional in CP2K through GauXC<\/a>.&#8221;<\/p>\n\n\n\n<h2 id=\"a-living-performance-report-for-skala\" class=\"wp-block-heading\">A living performance report for Skala<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Accuracy and broad availability only translate into scientific impact if Skala is also fast. Today, Skala can deliver performance comparable to semi-local meta-GGAs on both CPU\u2019s (with an overhead that disappears for molecules with more than 20-30 atoms) and GPUs, and we are committed to preserving that efficiency as it is integrated across the electronic-structure software ecosystem.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">But performance is not a fixed property. New Skala releases, improvements in libraries such as GauXC, and hardware-specific optimizations continuously improve efficiency and reveal new opportunities for further gains. Capturing this progress requires more than a single benchmark snapshot.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">To provide a transparent and up-to-date view of Skala&#8217;s performance, we are publishing a <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/aka.ms\/SkalaDFT\/timing\" type=\"link\" id=\"https:\/\/aka.ms\/SkalaDFT\/timing\" target=\"_blank\" rel=\"noopener noreferrer\">benchmarking harness together with a living performance report<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> that will be updated as new optimizations become available. This report tracks performance across a range of tasks and hardware platforms, while the harness enables package developers to benchmark, validate, and improve their own Skala implementations.<\/p>\n\n\n\n<figure class=\"wp-block-image aligncenter size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"2518\" height=\"1108\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1.png\" alt=\"Fig. 3 (figure attached): Benchmark of the computational cost of Skala-1.1 relative to r2SCAN, B3LYP, and M06-2X on GPUs and CPUs. Skala-1.1 achieves performance comparable to r2SCAN on GPUs and approaches semilocal-functional cost on CPUs for larger systems, while remaining significantly less expensive than hybrid functionals.\" class=\"wp-image-1184172\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1.png 2518w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1-300x132.png 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1-1024x451.png 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1-768x338.png 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1-1536x676.png 1536w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1-2048x901.png 2048w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/fig3_SKALA1.1-240x106.png 240w\" sizes=\"auto, (max-width: 2518px) 100vw, 2518px\" \/><figcaption class=\"wp-element-caption\">Figure 3: Computational cost of Skala on GPU and CPU, compared with a popular metaGGA functional (r2SCAN) and two hybrid functionals (B3LYP and M06-2X). On GPU, Skala 1.1 has the same cost as r2SCAN, and the hybrid functionals become more expensive for systems with more than ~1000 orbitals. On CPU, Skala has an overhead with respect to the other functionals for smaller systems, that disappears for systems with more than ~300 orbitals. <\/figcaption><\/figure>\n\n\n\n<h2 id=\"acknowledgments\" class=\"wp-block-heading\">Acknowledgments<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Skala is the product of a truly collaborative effort across AI for Science, and we thank our engineering, project management, and business operations teams for making this work possible. We also thank MSR Accelerator for their partnership in advancing data generation efforts and accelerating software integrations that help bring Skala to the broader scientific community.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Skala 1.1, the updated deep-learning exchange-correlation functional from Microsoft Research, provides greater accuracy, expanded accessibility across the computational chemistry ecosystem, and a living benchmark to track computational performance. <\/p>\n","protected":false},"author":43868,"featured_media":1184230,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"user_nicename","value":"Sebastian Ehlert","user_id":"42804"},{"type":"user_nicename","value":"Stefano Battaglia","user_id":"44283"},{"type":"user_nicename","value":"Thijs Vogels","user_id":"43464"},{"type":"user_nicename","value":"Jan Hermann","user_id":"42468"},{"type":"user_nicename","value":"Jens Wehner","user_id":"44278"},{"type":"user_nicename","value":"Giulia Luise","user_id":"43903"},{"type":"user_nicename","value":"Klaas Giesbertz","user_id":"43906"},{"type":"user_nicename","value":"Chin-Wei Huang","user_id":"41533"},{"type":"user_nicename","value":"Aaron Kaplan","user_id":"44235"},{"type":"user_nicename","value":"Kate Milton","user_id":"44284"},{"type":"user_nicename","value":"Stephanie Marisa Lanius","user_id":"43905"},{"type":"user_nicename","value":"Derk Kooi","user_id":"42702"},{"type":"user_nicename","value":"P. Bern&aacute;t Szab&oacute;","user_id":"44079"},{"type":"user_nicename","value":"Gregor Simm","user_id":"41811"},{"type":"user_nicename","value":"Rianne van den Berg","user_id":"40852"},{"type":"user_nicename","value":"Paola Gori Giorgi","user_id":"41746"}],"msr_hide_image_in_river":0,"footnotes":""},"categories":[1],"tags":[],"research-area":[13556],"msr-region":[],"msr-event-type":[],"msr-locale":[268875],"msr-post-option":[243984],"msr-impact-theme":[],"msr-promo-type":[],"msr-podcast-series":[],"class_list":["post-1184167","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-research-blog","msr-research-area-artificial-intelligence","msr-locale-en_us","msr-post-option-blog-homepage-featured"],"msr_event_details":{"start":"","end":"","location":""},"podcast_url":"","podcast_episode":"","msr_research_lab":[851467],"msr_impact_theme":[],"related-publications":[],"related-downloads":[],"related-videos":[],"related-academic-programs":[],"related-groups":[],"related-projects":[1149500],"related-events":[],"related-researchers":[{"type":"user_nicename","value":"Sebastian Ehlert","user_id":42804,"display_name":"Sebastian Ehlert","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sehlert\/\" aria-label=\"Visit the profile page for Sebastian Ehlert\">Sebastian Ehlert<\/a>","is_active":false,"last_first":"Ehlert, Sebastian","people_section":0,"alias":"sehlert"},{"type":"user_nicename","value":"Stefano Battaglia","user_id":44283,"display_name":"Stefano Battaglia","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/sbattaglia\/\" aria-label=\"Visit the profile page for Stefano Battaglia\">Stefano Battaglia<\/a>","is_active":false,"last_first":"Battaglia, Stefano","people_section":0,"alias":"sbattaglia"},{"type":"user_nicename","value":"Thijs Vogels","user_id":43464,"display_name":"Thijs Vogels","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/thijsvogels\/\" aria-label=\"Visit the profile page for Thijs Vogels\">Thijs Vogels<\/a>","is_active":false,"last_first":"Vogels, Thijs","people_section":0,"alias":"thijsvogels"},{"type":"user_nicename","value":"Jan Hermann","user_id":42468,"display_name":"Jan Hermann","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/janhermann\/\" aria-label=\"Visit the profile page for Jan Hermann\">Jan Hermann<\/a>","is_active":false,"last_first":"Hermann, Jan","people_section":0,"alias":"janhermann"},{"type":"user_nicename","value":"Jens Wehner","user_id":44278,"display_name":"Jens Wehner","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/jenswehner\/\" aria-label=\"Visit the profile page for Jens Wehner\">Jens Wehner<\/a>","is_active":false,"last_first":"Wehner, Jens","people_section":0,"alias":"jenswehner"},{"type":"user_nicename","value":"Giulia Luise","user_id":43903,"display_name":"Giulia Luise","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/giulialuise\/\" aria-label=\"Visit the profile page for Giulia Luise\">Giulia Luise<\/a>","is_active":false,"last_first":"Luise, Giulia","people_section":0,"alias":"giulialuise"},{"type":"user_nicename","value":"Klaas Giesbertz","user_id":43906,"display_name":"Klaas Giesbertz","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/kgiesbertz\/\" aria-label=\"Visit the profile page for Klaas Giesbertz\">Klaas Giesbertz<\/a>","is_active":false,"last_first":"Giesbertz, Klaas","people_section":0,"alias":"kgiesbertz"},{"type":"user_nicename","value":"Chin-Wei Huang","user_id":41533,"display_name":"Chin-Wei Huang","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/chinweihuang\/\" aria-label=\"Visit the profile page for Chin-Wei Huang\">Chin-Wei Huang<\/a>","is_active":false,"last_first":"Huang, Chin-Wei","people_section":0,"alias":"chinweihuang"},{"type":"user_nicename","value":"Aaron Kaplan","user_id":44235,"display_name":"Aaron Kaplan","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/aaronkaplan\/\" aria-label=\"Visit the profile page for Aaron Kaplan\">Aaron Kaplan<\/a>","is_active":false,"last_first":"Kaplan, Aaron","people_section":0,"alias":"aaronkaplan"},{"type":"user_nicename","value":"Kate Milton","user_id":44284,"display_name":"Kate Milton","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/katemilton\/\" aria-label=\"Visit the profile page for Kate Milton\">Kate Milton<\/a>","is_active":false,"last_first":"Milton, Kate","people_section":0,"alias":"katemilton"},{"type":"user_nicename","value":"Stephanie Marisa Lanius","user_id":43905,"display_name":"Stephanie Marisa Lanius","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/slanius\/\" aria-label=\"Visit the profile page for Stephanie Marisa Lanius\">Stephanie Marisa Lanius<\/a>","is_active":false,"last_first":"Lanius, Stephanie Marisa","people_section":0,"alias":"slanius"},{"type":"user_nicename","value":"Derk Kooi","user_id":42702,"display_name":"Derk Kooi","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/derkkooi\/\" aria-label=\"Visit the profile page for Derk Kooi\">Derk Kooi<\/a>","is_active":false,"last_first":"Kooi, Derk","people_section":0,"alias":"derkkooi"},{"type":"user_nicename","value":"P. Bern&aacute;t Szab&oacute;","user_id":44079,"display_name":"P. Bern&aacute;t Szab&oacute;","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/bernatszabo\/\" aria-label=\"Visit the profile page for P. Bern&aacute;t Szab&oacute;\">P. Bern&aacute;t Szab&oacute;<\/a>","is_active":false,"last_first":"Szab\u00f3, P. Bern\u00e1t","people_section":0,"alias":"bernatszabo"},{"type":"user_nicename","value":"Gregor Simm","user_id":41811,"display_name":"Gregor Simm","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/gregorsimm\/\" aria-label=\"Visit the profile page for Gregor Simm\">Gregor Simm<\/a>","is_active":false,"last_first":"Simm, Gregor","people_section":0,"alias":"gregorsimm"},{"type":"user_nicename","value":"Rianne van den Berg","user_id":40852,"display_name":"Rianne van den Berg","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/rvandenberg\/\" aria-label=\"Visit the profile page for Rianne van den Berg\">Rianne van den Berg<\/a>","is_active":false,"last_first":"van den Berg, Rianne","people_section":0,"alias":"rvandenberg"},{"type":"user_nicename","value":"Paola Gori Giorgi","user_id":41746,"display_name":"Paola Gori Giorgi","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/pgorigiorgi\/\" aria-label=\"Visit the profile page for Paola Gori Giorgi\">Paola Gori Giorgi<\/a>","is_active":false,"last_first":"Gori Giorgi, Paola","people_section":0,"alias":"pgorigiorgi"}],"msr_type":"Post","featured_image_thumbnail":"<img width=\"960\" height=\"540\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-960x540.jpg\" class=\"img-object-cover\" alt=\"Schematic of the Skala architecture, showing how meta-GGA electronic features are transformed through point-wise processing and non-local atomic interactions to predict density functional theory energies.\" decoding=\"async\" loading=\"lazy\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1-1280x720.jpg 1280w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2026\/08\/Skala-BlogHeroFeature-1400x788-1.jpg 1400w\" sizes=\"auto, (max-width: 960px) 100vw, 960px\" \/>","byline":"","formattedDate":"August 20, 2026","formattedExcerpt":"Skala 1.1, the updated deep-learning exchange-correlation functional from Microsoft Research, provides greater accuracy, expanded accessibility across the computational chemistry ecosystem, and a living benchmark to track computational performance.","locale":{"slug":"en_us","name":"English","native":"","english":"English"},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1184167","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/users\/43868"}],"replies":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/comments?post=1184167"}],"version-history":[{"count":39,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1184167\/revisions"}],"predecessor-version":[{"id":1184319,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1184167\/revisions\/1184319"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/1184230"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1184167"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/categories?post=1184167"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/tags?post=1184167"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1184167"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=1184167"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=1184167"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1184167"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1184167"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1184167"},{"taxonomy":"msr-promo-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-promo-type?post=1184167"},{"taxonomy":"msr-podcast-series","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-podcast-series?post=1184167"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}