PyLumerical is a new simplified and modernized way to access the Lumerical Python API. Fully compatible with the PyAnsys ecosystem, you can use PyLumerical to interact with Lumerical tools (Ansys Lumerical FDTD™, Ansys Lumerical MODE™, Ansys Lumerical Multiphysics™, and Ansys Lumerical INTERCONNECT™) using the popular Python programming language, alongside other Ansys products such as PyOptislang, PyAEDT, PySpeos and more, and open-source libraries. The 2025 R2.3 release includes support for pip install, basic developer documentation, and a few practical pythonic examples to get started. You can now launch your favorite Python IDE and use PyLumerical within Python Libraries to script simulations, automate workflows, and integrate photonics design into broader multiphysics workflows. PyLumerical is available on GitHub and PyPI, and customers are welcome to contribute to the GitHub repository to help shape future API and share enhancements with the broader Lumerical photonics community.
Persisting Solve Checkouts for Enterprise Licensing: Lumerical introduces support for persisting solve checkouts when using an Enterprise license for Lumerical FDTD, Lumerical MODE, Lumerical Multiphysics and Lumerical INTERCONNECT. This enhancement ensures that licenses checked out during a simulation are retained across multiple simulation runs within the same session, reducing the overhead of repeated license requests and improving simulation efficiency in high-performance computing (HPC) and batch processing environments. This feature is beneficial for workflows involving sweeps or optimizations, where maintaining license continuity can significantly streamline execution and reduce latency.
Ansys Optics Launcher: The performance of the Ansys Optics Launcher is improved. The launcher now loads images representing the examples faster and launches with faster start up times. You can also now change the location and size of the cache for Lumerical products.
Ansys Lumerical FDTDTM
Inverse Design: A new PortTransmission class is available in Lumopt, similar to the existing ModeMatch class. This enhancement will allow Lumopt optimizations to use FDTD ports directly, eliminating the need for users to place DFT monitors and configure ModeMatch when interested in waveguide transmission as Figure of Merit.
FDTD solver memory: The memory usage during meshing for FDTD simulations with PEC materials that uses the “snap to PEC” option has been reduced. In our benchmarks, we observed a 40% decrease in peak memory usage during meshing with similar total meshing time.
RCWA solver: The RCWA solver has undergone some improvements and got new features in 2025 R2.3 release:
Accuracy & Propagation Improvements
Corrected Field Monitor Output: Fixed phase errors in the fields, as well as field errors when using a non-orthogonal unit cell
Improved Incident Field Representation: Accurate source field before the first interface.
Backward Propagation Definition: Introduced selectable backward propagation mode, which can be “mirror k vector” or “reverse k vector” (default: “mirror k vector”).
New Features
Theta and Phi 2D map: Introduced a new option to split incident angles theta and phi into separate parameters when they are returned in the RCWA object and RCWA field monitors for improved plotting and scripting.
Field results outside of solver region: You can now visualize field results outside of the RCWA solver region. An option to clip fields to the solver region is added to the RCWA field monitor and is on by default.
View the index profile directly from field monitors: You can view the index on the same grid as the field monitor before or after the RCWA simulation.
View the grid using field monitors: This makes it easy to add a structure outline to the field images in the Visualizer, in the same way that it can be done with FDTD field monitors.
Ansys Lumerical INTERCONNECTTM
Simulation performance speed: We improved Lumerical INTERCONNECT simulation speed over 2025 R2.2 and 2025 R2.3 releases, especially for large-scale or long-duration circuits with high sample rates, in sample mode. Users working with Transmitter and Dispersion Eye Closure Quaternary (TDECQ), eye diagram, and Traveling Wave Laser Model (TWLM) elements, or hundreds of interconnected elements will benefit reduced simulation times and lower memory usage. In benchmark tests, noticeable speeds-up were observed across all large-scale simulations, ranging from 2X for distributed Bragg reflector (DBR) TWLM, to 16x for Ring Modulator bit error rate (BER) analysis.
Ansys Lumerical Multiphysics™
Thread settings: Multithread settings have been moved from the object properties of each solver into the resource configuration window. If you had configured multithread settings for previous projects, you need to reconfigure them in the resource configuration window.
Ansys Lumerical CML Compiler™
Automated model data collection: Four new automated data collection wizards have been introduced as a part of the CML Compiler automated model data collection tools. You can use these GUI-based tools to process simulations and rapidly generate compiled compact models for waveguides, s-parameter (fixed), s-parameter (parameterized), and thermal phase shifters. The wizards allow you to generate them directly for use in Lumerical INTERCONNECT or add them to an existing compact model library to be compiled using CML Compiler. The wizards only require a small number of inputs, streamlining compact model generation when simulation results are already available.