Let's define "boot time" as: time it takes from power coming on (which may be when the laptop comes out of sleep) to useful functionality, which may be serving network packets, having a shell login prompt, or a desktop environment responding to keyboard and mouse.
Single most important factor: Is the machine actually booting (re-initializing the hardware, OS kernel and all services), or is it just restoring from a hibernate or sleep state. May OSes (in particular MacOS and Windows) do not actually boot when you open a laptop or turn it on. So then the question becomes: How to enable sleep/hibernate when using FreeBSD on that particular hardware.
Second factor: Disk IO. With SSDs (including NVMe) as boot disks, one can get the boot time to around 10 seconds and below. With spinning disks, it's hard to get below 20-30 seconds (can be way more if lots of services are enabled).
Third factor: Hardware initialization, which mostly is BIOS, and (as SirDice already said) initialization of disk controllers. On a large enterprise-grade server with hundreds of (spinning) disks attached, this can take 10-15 minutes, because all those disks have to be spun up.
The OP is talking about using different init systems. My intuition is that this is a fool's errand, and won't make a significant difference. Similarly, creating a custom kernel (for example without modules, and only the needed drivers compiled in) no longer makes a significant difference.