The link provided by @balaji.bandi is a very helpful discussion of the best path selection in BGP (and a well deserved +5 for that contribution). The original post asks this question "Does that mean it's the best in the BGP table or it's the best in the Routing table?" Both previous responses have stressed that BGP best path is determined by comparing candidate routes in BGP and selecting the best candidate path. But they do not really address the comparison to the routing table.
It might be helpful to remember that there is one process in BGP which evaluates candidate routes and selects the best path. And that there is another process which manages inserting routes into the IP routing table. So the BGP process evaluates the routes and selects the best path. BGP then sends the route to the route table maintenance process which decides whether the route will be inserted into the routing table. There are some circumstances where the BGP route will not be inserted into the routing table. The easy example is if there is a static route for the same prefix (and with the default AD for static routes). In this case the route table maintenance process informs BGP that the route is not inserted into the routing table. And BGP marks that route as "rib failure" shows as r in the output of show ip bgp. This link has a discussion of rib failure and an example showing output of show ip bgp which has 2 prefixes with the r for rib failure and also has the > best path indicator.
https://www.cisco.com/c/en/us/support/docs/ip/border-gateway-protocol-bgp/213286-understand-bgp-rib-failure-and-the-bgp-s.html
So the answer to the original question is that the BGP best path prefix is the best prefix in the BGP table but not necessarily the best prefix in the ip routing table.
HTH
Rick
