Which cooler are you using? The added airflow and pressure from the extra 1,000 rpm won't significantly impact temperatures for most coolers.
As for noise equivalence, I alluded to that briefly above when I wrote "They [the 3,000 rpm fans] are annoying even if you turn down the speed in software or BIOS because they will spin up when you restart and boot," and also when you wake the machine from sleep. Both fan models are identical besides the speed, and motor noise isn't an issue on these fans. So the only noise variable is the rotational speed (i.e. 2,000 vs. 3,000). That said, you're always choose the 3,000 rpm version. I didn't.
I personally asked myself whether I foresaw any reasonable use for the fan that would benefit from spinning up the fan to more than 2,000 rpm. Luckily there wasn't. If there had been, I would have been listening to the fans ramp up every time I wake the computer from sleep, restart it, or boot it.
To find that out, I tested temperatures using the GPU and CPU meters, as well as the HDD, SSD, and motherboard read outs. I also have 4 temperature probes placed in key airflow locations in my case, in addition to having the ambient temperatures. Running Prime 95 and Furmark, nothing came close to a temperature of concern even when the fans are spinning at 600 rpm. Testing different rpms, I also found that the fans aren't audible until you spin up to 1,000+ rpms. So for me it turned out that even the 2,000 rpm version left about 60% overhead. On my setup, that meant that adding any additional overhead with a faster fan would have been worse than useless because of the unwanted noise.
Keep in mind that whether this applies to you will depend on your setup's cooling, components, airflow, and cleanliness. But if you have those things under control, there you'll likely be in a similar situation.