Epic said its preview launcher had a 46% cold-start advantage over Steam. A clip posted the next day shows Steam opening in 5.23 seconds, but the update's development method remains undisclosed.
Epic's comparison hinged on how long each client took to reach its store page. On October 14, 2026, the company released a preview of a rebuilt Epic Games Launcher for Windows and macOS. Linux support arrived for the first time, and startup time took centre stage in Epic's comparison of the two PC storefronts.
Epic sets the benchmark
Epic said its launcher reached the store page in an average of 5.47 seconds from a cold start, compared with 10.07 seconds for Steam. That amounted to a claimed 46% advantage. In a separate test, restoring the app from the System Tray to the Library took 0.36 seconds for Epic's launcher and 0.42 seconds for Steam, or about 14% faster. A cold start begins after the app has been fully closed; tray restoration returns an app that is already running in the background.
VideoCardz reported that independent tester Keywords Studios ran each system five times and excluded the three weakest systems from its final averages. Epic called the client an early preview, describing it as "light on features" and likening it to a "speedboat and not a yacht." PC Gamer described the release as a complete rebuild of the client.
Valve's Steam Client Update arrived on October 13, a day before Epic's preview. Its listed changes included improved startup performance, along with changes to Big Picture Mode and fixes. The timing drew attention, but Valve's update notes did not say the work was a response to Epic's comparison.
A faster Steam launch
On October 15, 2026, @girlglock_ shared a video on X showing Steam open in 5.23 seconds. That is 4.84 seconds faster than the 10.07-second Steam result in Epic's comparison. It is also 0.24 seconds below Epic's reported 5.47-second cold-start average.
The clip shows one launch, not a documented head-to-head benchmark. It gives no test setup or repeat-run data. The conditions are not established as identical to those in Keywords Studios' tests. The video captures a faster Steam launch in that particular run, but it does not offer a controlled comparison of both clients.
Epic's published figures put its preview ahead in the original cold-start comparison. The later clip records Steam launching below Epic's stated average. Its 5.23-second result is one shown run, not a new Steam average or a measurement of both launchers on the same system.
The coding method is unknown
Epic described its preview launcher as "vibe-coded," a reference to an AI-assisted development approach. The method behind the Steam Client Update has not been specified. Startup times alone do not identify which coding approach was used.
Epic tied the rebuild to faster navigation between the store and library. It also cited faster startup and reduced background load, alongside improvements to the library and downloads. The company presented the release as a preview, not a final version.
The 5.23-second Steam launch is the result shown in the video. The development method behind the update remains unknown, and the conditions of that run are separate from the startup-time measurement.
A cold start measures the time an application takes to open from a stopped state and load the store page. Restoring from the System Tray measures a different action because the application may already be running in the background. Epic's reported tray results differ by 0.06 seconds: 0.36 seconds for its launcher and 0.42 seconds for Steam. Those figures should not be combined with the cold-start measurements. The timings cover launch responsiveness, not the full experience of using a storefront. They also do not measure performance across every PC.
