RTX 4070 vs RTX 4080 - DON’T Buy a 4070 Gaming Laptop!

Nvidia RTX 4070 gaming laptops are not worth  buying compared to the higher tier RTX 4080. Yeah, a gaming laptop with a 4080  inside is going to cost more money, but it performs so much better in most games  that it’s worth spending the extra money. I’ve compared both GPUs in 25 games at 4K, 1440p and 1080p resolutions to show you  how crazy the difference is! This is critical information that anyone considering  an RTX 4070 laptop needs to pay attention to! But first, Gigabyte have  sponsored t

his part of the video. Gigabyte's high-end Aorus gaming  laptops have been redesigned this year, while budget conscious gamers  are covered by the updated G5. These laptops are more powerful than ever with  Nvidia's latest GeForce RTX 40 series graphics, allowing you to enhance your gaming  experience and get smoother gameplay with DLSS 3 frame generation in the latest titles. And Gigabyte have got content creators  covered with their newly updated Aero 16 and brand new Aero 14 for ultimate port

ability. Check out the sponsored  link below to find out more. Back to the comparison. The higher tier  RTX 4080 laptop GPU has 61% more CUDA, tensor and ray tracing cores compared to the lower  tier RTX 4070 - a massive difference. The 4080 also has 50% more VRAM capacity with a bigger  memory bus, more memory bandwidth, and faster memory. The 4080 can also hit higher boost clock  speeds, as it’s got a higher maximum power limit. In theory, both of these laptops can run with  an extra 25 watts

with Nvidia’s dynamic boost, but in the real world with an actual game running, 4070 laptops hit a voltage limit in  most games at around 100 watts or so, which means there’s no FPS boost beyond this  point as it can’t actually run above 100 watts. This is not a problem for the 4080, and FPS in  games continues to scale at higher power levels. This graph shows the 3DMark Time Spy graphics  score from each GPU at different power levels. The 4070 can run lower, making it better  for thinner and li

ghter gaming laptops, while the 4080 can use more power, which is  why it can be found in bigger laptops with more cooling. With both GPUs running at 60 watts,  the 4080 was scoring 20% higher than the 4070, but if we limit both at 140 watts,  well the 4080 has a much larger 49% lead. And then it gets even further  ahead at its maximum 175 watt limit. I’m comparing both of these GPUs at their max, so we are expecting a pretty big performance  difference. So expect a smaller change to what we’re

going to see here in smaller  laptops with lower power limits. I’m using XMG’s Neo 16 laptop to do this  testing, so we’ve got the exact same CPU, RAM, SSD and water cooling for  a fair comparison. They’ve both got the same sized batteries  too, and with a game running, the 4070 laptop lasted 20% longer with both on  battery power and limited to a 30 FPS frame cap. The 4070 laptop uses less power. When measuring  both at the wall with the charger connected, the 4080 was drawing 19% more power  i

n Control at 4K high settings, and 24% more power in Cyberpunk  2077 at 4K ultra settings. Yeah, the 4080 offers more  FPS, as you’ll see soon, but from a performance per watt  perspective, the 4070 ends up more power efficient. The 4080 hits higher frame  rates, but it needs more power to do so. The 4070 was also running a little cooler in these  games too. The 4080 uses a physically larger die, so in theory it has more surface area for cooling, but again it’s using around 20% more power  too,

and more power results in more heat. The clock speed difference wasn’t that big, with the 4070 actually clocking higher in  Control. The 4080 has 61% more CUDA cores, but with 19% more power it doesn’t seem  that it can clock them quite as high. With all that in mind, let’s move into our 25 game  comparison, starting with Cyberpunk 2077. I’ve got the 1080p results down the bottom, 1440p in the  middle, and 4K up the top, with the 4070 results underneath the 4080 results. Despite the small  diffe

rence seen in the clock speeds, the 4080’s 61% more CUDA cores are clearly being put to work  here, as it’s reaching a 64% higher average frame rate at 4K, though the 1080p difference  was smaller with a 26% lead over the 4070. Some games like Apex Legends had massive  performance differences at 1080p though. The 4080 was 65% faster than the 4070, the  biggest difference at this lower resolution out of all 25 games tested. Even the dips in  performance on the 4080, as shown by the 1% lows, were

higher than the average FPS coming  out of the 4070 - a big difference. Then on the other hand, games that are more CPU  heavy like Microsoft Flight Simulator barely had much difference at all at 1080p. There’s  a bigger 36% lead with the 4080 at 1440p, but this is still the smallest  difference out of our 25 game selection at this resolution. By the time  we get up to 4K, the 4080 was 54% faster, because the more powerful GPU can better  stretch its legs when we’re more GPU bound. Now although

the 4070 has done relatively alright  at 4K so far, it starts to fall apart in Far Cry 6. Its 8 gigs of VRAM just isn’t enough at this  resolution with ultra settings and HD textures on, though you could of course lower the settings to  reach a higher frame rate. My point here is that the extra 4 gigs of VRAM available to the 4080 do  help it out in some games at higher resolutions. This is also seen in Forza Horizon 5 with the  highest extreme setting preset. At 4K the 4080 has a massive 92% hi

gher average frame  rate - almost double the FPS of the 4070, crazy stuff. At 1440p the 4080 was  49% faster, which sounds great, but it’s actually a below average  result out of the 25 games tested. Watch Dogs Legion can be fairly  VRAM heavy at ultra settings too, which would explain why the 4080 was 77% faster  at 4K, one of the biggest differences from the 25 games tested. Even at 1440p and 1080p,  the 1% lows from the 4080 were still better than the average FPS from the 4070, which just  me

ans the 4080 is more stable and consistent. It’s a similar deal in Dying Light 2 as well,  the 4080 has a 77% lead over the 4070 at 4K, but then on the other hand, there’s almost no  difference in average FPS in The Witcher 3 at 4K with the ray tracing preset enabled. This is  a weird result that I’ve seen with a number of laptops in this game, with ray tracing on once  it gets too hard to run it seems to just cap out at 30 FPS. Even 1080p was 59% faster on the 4080,  the third biggest differenc

e at this resolution. Metro Exodus also enables some RT  effects with the highest setting preset, which probably explains why we’re seeing  the same 59% improvement with the 4080 at 1080p. Tracing those rays just needs more GPU  horsepower, even at this lower resolution. The differences in Hogwarts Legacy were below  average compared to most other games tested, and although the 4070 was only reaching 30 FPS  at 4K max settings, don’t forget that both GPUs could see a nice improvement by making u

se  of features like DLSS and frame generation. Alright after 10 games I think you’ve got the  general idea, the 4080 is significantly better in most cases. I’ll just quickly skip through  the rest of the 15 games tested on screen now, instead of wasting your time talking through the  results from each individual game. I think it’s important to test a wide selection of games  so that we can get an accurate picture of the average performance differences to make the  fairest possible conclusion. A

s I always say, more data equals more better. Let’s  look at those average differences next. On average over all 25 games tested, at  1080p the higher tier RTX 4080 laptop GPU was 44% faster compared to the lower tier  RTX 4070 laptop GPU. This graph shows how much faster each game was on the 4080,  and as you can see, results really vary depending on the specific game. CPU heavier  games like Flight Simulator and Spider Man were 5 to 10% faster on the 4080, while a  number of others were closer

to 60% faster. If we step up to the higher 1440p  resolution where we’re more GPU bound, the RTX 4080 was now 55% faster than the 4070 on  average. This time even the smallest difference was 36% faster with the 4080, while in the best  case, we’re seeing up to a 70% lead in Dead Space. The gap gets bigger at the higher 4K resolution,  where we’re even more GPU dependent, and the VRAM difference starts to matter more in some games.  The 4080 is now 69% faster than the 4070, nice, although if we

treat that massive Far  Cry 6 result at the top as an outlier, then the 4080 was instead 59% ahead on average. Here’s how frame rates look if we instead  take the average of all 25 games. I think this better allows us to visually see the overall  differences as a quick summary. It lets us easily see that even over a wide selection of games, the  4080 has a big performance uplift over the 4070. Alright, so the 4080 performs  way better compared to the 4070, but is it worth spending  more money to

get the 4080? Prices of both will change over time, so refer to the links below the  video for updates and current sales. And if any 4070 or 4080 gaming laptops do  go on sale, we’ll be sure to add them to our gaminglaptop.deals website. We  update that every day with all of the latest sales so you can save  money on your next gaming laptop! Alright, I’ve got a few examples. The  ASUS ROG Strix G18 costs $2300 USD for the RTX 4070 model, while the 4080  is only $200 more, or 9% more money. Raze

r’s Blade 16 certainly costs  a whole lot more money overall, but it’s still only 9% more money to  go for the better performing 4080. If we actually use XMG’s Neo 16 with liquid  cooler that I’ve tested in this video, we’re looking at 25% more money to go from the  4070 up to the 4080. This same model is also sold in the US by Eluktronics, and we’re looking  at a similar 21% extra money for the 4080. The 4080 ends up way better from a cost  per frame perspective compared to the 4070 in all thre

e of these examples, so for  the ASUS ROG Strix G18, Razer Blade 16, and XMG Neo 16 / Eluktronics Mech-16 GP. Yeah,  the 4080 costs a few hundred dollars more, but it performs so much better in games that it  ends up being worth it by offering better value. It’s a bit different with something like  the ASUS Zephyrus M16. There’s a much bigger $750 price gap between the 4070 and 4080  options, so 38% more money for the 4080, and unlike the last three laptops,  the M16 is a thinner design, so the

4080 is power limited to 120 watts plus  dynamic boost. This means it won’t perform as well as the 4080 results shown previously in  this video, so I can’t fairly do the cost per frame calculation with my data, but it’s an  example where the 4080 may make less sense. Based on the big performance  differences between the 4070 and 4080, it kind of feels like Nvidia have just  left a space for a 4070 Ti next year. But that’s just me speculating based on  the big differences between these two. Hones

tly, I don’t really think that a 4070  laptop should be considered if your goal is 4K gaming. To be fair, as we saw in many of  the games, you can certainly get away with it. Especially if you’re willing to turn down the  settings a bit or make use of features like FSR, DLSS or frame generation. But ultimately  for 4K gaming with higher setting levels, I’d recommend the 4080 for its extra power,  but also the 50% extra VRAM capacity. Even for 1440p gaming, spending  a few hundred dollars extra i

f you can afford it for the 4080 just  feels like a no-brainer. Which kind of makes me wonder if the 4070 only exists  to entice people to spend more on the 4080. Now to be fair, the 4070 does make more  sense in thinner and lighter laptops as that GPU is pretty efficient at lower  power levels compared to last gen. But when we’re talking about max powered  options for the best FPS in games, the 4080 wins hands down and is  worth the extra money in most cases. But if you’re going to spend RTX 40

80 money, is it worth spending more and going all the  way up to the RTX 4090? Check out this video next where I’ve compared those two GPUs  in the same 25 games at 3 resolutions!