I would use the number from the actual test. Looking at my "estimated FTP" for the last 6 weeks on my Strava Power Curve, it's exactly equal to my best 20-minute power for that same period. I know from experience my real FTP is more like 93% of 20-minute power, so I'm going to say take the Strava number with a grain of salt. Granted in OP's case 6 watt difference is not much, but I see no reason to do an FTP test unless you're going to use the result to set your FTP.