Sony Boosts PS5 Pro Performance with PSSR 2.0 AI-Upscaling U

Sony has released a significant system software update for the PS5 Pro, introducing PSSR 2.0 AI-upscaling to the high-end console. This version of PlayStation Spectral Super Resolution arrives this week as part of the May 2026 firmware rollout, aiming to improve visual fidelity and performance across the platform's library.

The PSSR 2.0 AI-upscaling technology is the result of a development partnership between Sony and AMD, previously referred to as Project Amethyst. This iteration utilizes a rebuilt neural network designed to deliver sharper image quality while minimizing the visual artifacts that sometimes appeared in the initial launch version. By refining how the console reconstructs lower-resolution frames, the update provides a cleaner look for demanding titles.

Machine Learning Frame Generation and Performance

A major addition in this update is the inclusion of machine learning-based frame generation at the system level. This feature is intended to help the PS5 Pro maintain a consistent 60 frames per second in technically intensive open-world environments. Sony expects this capability to be particularly beneficial for upcoming major releases, such as the highly anticipated Grand Theft Auto VI, which is currently slated for a late 2026 arrival.

For players who enjoy older games, the firmware includes a new system-level toggle labeled Improve PSSR Image Quality. This setting allows the console to apply its advanced upscaling techniques to backward-compatible titles and existing PS5 games that have not received a specific Pro patch. Users can find this option within the system's screen and video settings menu.

The update is available for download immediately for all PS5 Pro owners. While the primary focus is on the new PSSR 2.0 AI-upscaling and frame generation tools, the software package also includes standard stability improvements and minor interface refinements to the PlayStation 5 operating system.

While we strive for accuracy, bytevyte can make mistakes. Users are advised to verify all information independently. We accept no liability for errors or omissions.

Photo by Georgiy Lyamin on Unsplash

✔Human Verified

Machine Learning Frame Generation and Performance

Related Articles