Vercel Flags starts 7x faster and no longer extends function duration
Vercel Flags 1.5.3 cuts internal-test p75 cold-start initialization from over 70 ms to 10 ms and stops usage reporting from extending function duration.
The SDK now reads bundled flag data and checks its revision without a network request. In internal testing, p75 cold-start initialization fell from **over 70 ms to 10 ms**; p99 flag reads are **under 1 ms**. Usage reporting no longer keeps a function running after its response.
If you use Vercel Flags, update `@flags-sdk/vercel` to **1.5.3 or later** and redeploy to get the changes.
The SDK now reads bundled flag data and checks its revision without a network request. In internal testing, p75 cold-start initialization fell from **over 70 ms to 10 ms**; p99 flag reads are **under 1 ms**. Usage reporting no longer keeps a function running after its response. If you use Vercel Flags, update `@flags-sdk/vercel` to **1.5.3 or later** and redeploy to get the changes. The initialization figures come from internal testing; the material does not give a workload or methodology for comparing them with other setups.
This reports lower Flags initialization latency from bundled data and revision checks, plus usage reporting that no longer extends function duration. The p75 figures are internal; the source gives no workload or cross-setup comparison, so the gains are specific to the stated SDK behavior.