Starting with Windows 11's cumulative update "KB5121003" released on August 11, 2026, the Low Latency Profile (LLP)—which briefly raises CPU frequency—has begun working during the launch of general applications as well. Microsoft's official update history doesn't mention the LLP name or its expanded scope, but Windows Latest compared the same PC before and after the update and confirmed the feature activating across multiple apps. This means the improvements, which had previously centered on the Windows Shell like the Start menu and Search, have now spread to everyday software people use. However, this isn't an update that arrives simultaneously on all PCs, nor is there any guarantee that every app will see the same speed improvements.
Short-burst boost expands to general apps
KB5121003 targets all editions of Windows 11 24H2 and 25H2, updating the OS Build to 26100.9168 and 26200.9168 respectively. Beyond the August security fixes, it's a cumulative update that also includes previously distributed improvements, such as the July 28 preview update "KB5101684." Microsoft announced in the May 26 "KB5089573" release that it would speed up app launches. The same release notes also mentioned improvements to the Start menu, Search, and Action Center, but didn't explain the underlying mechanism.
This latest change became clearer through Windows Latest's before-and-after testing. The outlet launched Notepad, Calculator, and Chrome in sequence on the same PC. It also tried Epic Games Store, Weather, PowerPoint, and VLC, reporting that the intentional frequency spikes—absent before the update—occurred in each app after applying KB5121003. The fact that this worked across native apps, WebView2 apps, and traditional Win32 software marks a difference from the earlier stage, which was limited to operations within the Shell.
On the other hand, what Microsoft's KB page definitively states is the update's target versions, Build numbers, and release date. While you can trace which past fixes are included, there's no explanation that the scope of LLP was expanded to general apps in August. Therefore, this expansion to general apps should be understood not as an official Microsoft announcement, but as a hands-on verification tracking behavior observed on a publicly released Build.
What the 1-3 second high-frequency operation aims to achieve
What LLP changes isn't sustained performance where the CPU runs at full capacity for extended periods. According to the mechanism reported by Windows Central in May, citing Microsoft insiders, the CPU is briefly pushed to maximum frequency for just 1-3 seconds in sync with high-priority operations like opening apps or menus. This follows the "race to sleep" philosophy: quickly finish the heavy processing right after launch, then return to a low-power state as soon as possible.
What needs to be distinguished here is frequency versus utilization. Just because something runs at high frequency doesn't mean all cores are under 100% load. What Windows Latest observed on a machine with an Intel Core i5-13420H wasn't a spike in utilization, but a brief behavior where the four P-cores rose to around 4.6GHz and the four E-cores exceeded 2GHz. Once processing finishes, the frequency drops back down, so LLP serves a different role than features designed to boost long-duration benchmark scores.
In Windows Latest's testing, frequency increases and faster launches were observed across apps built on different UI frameworks. However, it remains unclear how effective LLP is when non-CPU processing accounts for most of the wait time. Further testing across a wider range of apps and PC configurations, with quantitative measurement of launch times, is needed.
Can the "up to 40%" figure hold up in the stable release?
Regarding LLP, figures claiming app launches could speed up by as much as 40% and operations like the Start menu by as much as 70% circulated earlier. These originated from experimental test values reported by Windows Central on May 7, while the feature was still under development. The article at the time also noted that the feature was in early testing within the Windows Insider Program, and that the boost duration and frequency could change before general availability.
Windows Latest's testing in August confirmed that the same mechanism now works for general apps on a publicly available Build. On the same PC with an Intel Core i5-13420H, apps including Epic Games Store reportedly opened noticeably faster after the update. However, the article doesn't include a table showing the number of repeated trials, launch times in seconds for each app, or variance data. While the frequency changes in the P-cores and E-cores serve as evidence that the feature is active, they can't simply be translated into an effect size applicable to all users.
For this reason, the 40% and 70% figures are useful as reference points for understanding the scale LLP is aiming for, but they aren't performance guarantees for KB5121003. Launch times vary depending on CPU generation and cooling conditions. They're also affected by power mode, storage, and which apps are running in the background. As for heat generation and battery life, the design goal of quickly returning to idle is one thing—but confirming that the impact is negligible across a wide range of real-world devices is a separate matter that still needs verification.
Not a simultaneous rollout—arrival timing varies via CFR
Windows Latest reports that Controlled Feature Rollout (CFR) is being used to enable LLP for apps. According to Microsoft, CFR is a mechanism that gradually enables features starting with a subset of devices, expanding the rollout as readiness is confirmed. Even after installing KB5121003, there's no guarantee the feature will activate on all 24H2/25H2 devices on the same day.
As of Windows Central's May report, no user-facing toggle for LLP had been confirmed. While third-party tools exist that can forcibly enable the feature using its feature ID, waiting for the CFR rollout is the safer approach for typical use. For enterprise IT administrators as well, since the date an update is deployed and the date the feature actually activates can differ, it would be premature to judge the organization-wide impact based on just a handful of early devices.
Also worth noting is caution around the phrase "for all of Windows 11." KB5121003 directly targets 24H2 and 25H2, while 26H1 receives a separate cumulative update. While the significance of LLP reaching general apps is substantial, solidifying the assessment requires comparing device-specific launch times and power consumption—including battery-powered use—once CFR has fully propagated. If the reduction in wait time can be reproduced there, Windows 11's responsiveness improvements will shift from something felt subjectively to a measurable, verifiable outcome.