There should be no 10.139.* in /etc/resolv.conf: these addresses are only internal to Qubes and will be resolved by the net VM according to its configuration (e.g. your 192.168.18.1). This is the purpose of match in proto { udp tcp } from xnf0:network ... and pass in proto... in the pf.conf guide.
Your resolv.conf will be correct with only the first line (or automatically generated by dhclient?), and it seems to resolve correctly from the tcpdump traces in your previous post.
In the last image, 10.137.0.27 is sending ICMP requests, but there is no response. As it seems, from the 5th image in your previous post, there is also traffic on re0, so IP forwarding is correctly enabled. It also appears that the NAT process is correct as the outgoing packets have the IP address of re0 (192.168.18.109) as their source address. So the packets seem to be lost on the way back, either because they are not NATed back or because they are filtered. I don’t know enough about OpenBSD to track deeper on this issue 
