If you want to get a baseline of what your able to get from your Internet circuit, I would pull the router and wire in directly to the modem to do a speed test. This will eliminate any variables such as wireless interference, your router, or someone on your home network accessing the Internet.
With regards to if the speeds are realistic, yes they are up to a point. For residential service, it's always best effort for the ISP. They're not going to guarantee the performance of the line nor provide any specific SLA (service level agreements) around what constitutes a breach of the service contract. For this type of guarantee, you'll have to step up to a business class service with the requisite increase in monthly charges. Also whether a particular speed tier is worth it or not depends on how you use your Internet circuit. For me, I have Comcast. I have their 50 Mbps down service. I didn't get this because I need 50 down. Why I got it was due to the upload speeds which no one really focuses on. Because you're not only pulling data from the Internet, you're sending data up to it too. For the type of usage I have, upload speeds are more critical than download speeds as I host a ton of different services on my home network I access over the Internet. The Comcast 50 down service provides 10 Mbps up. Their next tier down at 25 Mbps only does 2 Mbps up. Verizon FIOS has been pointing out this asymmetric network speed in their "Half Fast" Internet campaign.
On the topic of routers, a common mistake many people think is because a particular router has a 100 Mbps or 1 Gig interface it must be forwarding traffic at that speed. This is farther from the truth. The performance of a particular router is dependent on the type of traffic being routed, the number of sessions being supported, and the type/number of services running on the router/firewall. I have yet to see any manufacturer publish actual performance data on their SOHO products. On the SMB side of the house, you do see some published performance specs. Take Netgear for example, there is no mention about how their SOHO routers perform but they do publish some information on their ProSafe SMB line. Not many review sites actually do a proper full blown performance test of SOHO routers. The closest is the site...smallnetbuilder.com where they use IXIA's ixcharriot. But I still have issues with their testing methodology. While they use a proper performance measuring tool, they don't do it properly. Their performance results are only doing a pure layer 3 route test. They turn off the SPI firewall and if applicable the NAT functions. While NAT shouldn't present much if any load on a router, the SPI sure will. This is why my SonicWall has different performance stats reported by the manufacturer. If I just run it as a DPI firewall, it'll do x speed. If I turn on the IPS/IDS function, it'll go slower and do x speed. If I turn on the UTM function, it will go even slower and do x speed. But most SOHO routers should be able to handle the speeds of the mainstream Internet plans.
Regarding your question about ac or not ac, it depends on your usage and whether you want to do a little future proofing. If you do a lot of internal file sharing of large files for example, then it would be advantageous to go ac. One clear advantage with going ac is that it operates strictly on the 5 GHz band. So you won't have as much issue with neighboring RF interference. In addition with ac, you have more channel selections to get your network away from neighboring network operating on 5 GHz. You can't say that for 2.4 GHz where you're limited to channels 1, 6, and 11 for non overlapping channels. The good thing about ac is the devices which support it are backwards compatible to n.
With regards to if the speeds are realistic, yes they are up to a point. For residential service, it's always best effort for the ISP. They're not going to guarantee the performance of the line nor provide any specific SLA (service level agreements) around what constitutes a breach of the service contract. For this type of guarantee, you'll have to step up to a business class service with the requisite increase in monthly charges. Also whether a particular speed tier is worth it or not depends on how you use your Internet circuit. For me, I have Comcast. I have their 50 Mbps down service. I didn't get this because I need 50 down. Why I got it was due to the upload speeds which no one really focuses on. Because you're not only pulling data from the Internet, you're sending data up to it too. For the type of usage I have, upload speeds are more critical than download speeds as I host a ton of different services on my home network I access over the Internet. The Comcast 50 down service provides 10 Mbps up. Their next tier down at 25 Mbps only does 2 Mbps up. Verizon FIOS has been pointing out this asymmetric network speed in their "Half Fast" Internet campaign.
On the topic of routers, a common mistake many people think is because a particular router has a 100 Mbps or 1 Gig interface it must be forwarding traffic at that speed. This is farther from the truth. The performance of a particular router is dependent on the type of traffic being routed, the number of sessions being supported, and the type/number of services running on the router/firewall. I have yet to see any manufacturer publish actual performance data on their SOHO products. On the SMB side of the house, you do see some published performance specs. Take Netgear for example, there is no mention about how their SOHO routers perform but they do publish some information on their ProSafe SMB line. Not many review sites actually do a proper full blown performance test of SOHO routers. The closest is the site...smallnetbuilder.com where they use IXIA's ixcharriot. But I still have issues with their testing methodology. While they use a proper performance measuring tool, they don't do it properly. Their performance results are only doing a pure layer 3 route test. They turn off the SPI firewall and if applicable the NAT functions. While NAT shouldn't present much if any load on a router, the SPI sure will. This is why my SonicWall has different performance stats reported by the manufacturer. If I just run it as a DPI firewall, it'll do x speed. If I turn on the IPS/IDS function, it'll go slower and do x speed. If I turn on the UTM function, it will go even slower and do x speed. But most SOHO routers should be able to handle the speeds of the mainstream Internet plans.
Regarding your question about ac or not ac, it depends on your usage and whether you want to do a little future proofing. If you do a lot of internal file sharing of large files for example, then it would be advantageous to go ac. One clear advantage with going ac is that it operates strictly on the 5 GHz band. So you won't have as much issue with neighboring RF interference. In addition with ac, you have more channel selections to get your network away from neighboring network operating on 5 GHz. You can't say that for 2.4 GHz where you're limited to channels 1, 6, and 11 for non overlapping channels. The good thing about ac is the devices which support it are backwards compatible to n.