Garmin under-reads distance, Strava gets it right

Not telling it is impossible, but am just curious how do you know that the distance in Garmin is wrong, and the "corrected" one in Strava correct? Couldn't it be the exact opposite? Did you try running for example on a stadium track to confirm the suspiction?

A fair question.  My main calibration was by using it over a long-established parkrun course, which has been accurately measured, and typically run by 1,000 people weekly. The fact that Strava takes the data points from the Instinct recorded every second and shows 5.00km as the distance, every week, when the Instinct typically shows something like 4.84km or 4.82km uncorrected, is pretty good evidence to me.  There is a track nearby that I could use, and I will throw the question to that parkrun group, but it's the watch that looks to be the problem; it seems unable to derive distance from the data the way Strava can.

OK, that sounds like a reasonabe argumentation. Consider writing the concern directly to Garmin, here on the community forum it risks not to get the attention of anyone from Garmin. You could try for example this form: https://www.garmin.com/en-US/forms/ideas/

Thanks, I'll try there.

It because i make the route with another 10 people that all of them reading 5k, and its because obvious that i dont take 25m to make 1k that why i know, that its not correct

Well, the standard accuracy of GPS is about 15m (source Wikipedia), and although it can be augmented with different methods, it is the case at dedicated expensive devices for military or professional use. And that is the accuracy for a single keypoint of the track. At a 1km section, there are typically several hundreds of keypoints, so the difference of 25m between two devices is nothing exceptional, and I'd tell a pretty good result.