Skip to main content

Better than 2080 Ti? RTX 3070 Benchmarks Hit Geekbench

RTX 3070
(Image credit: Nvidia)

The Nvidia GeForce RTX 3070 launches next Thursday, assuming you can get your hands on one, meaning that performance numbers are starting to leak. Up until now, we've had to rely on official charts from Nvidia, but the Geekbench Browser, which users can anonymously upload their benchmark results to, found itself with two new RTX 3070 results late last night. 

These tests were performed with Geekbench 4.0 rather than the newer Geekbench 5.0, but these numbers are still plenty relevant, and the use of an older benchmark actually makes it easier to compare these results with those of older GPUs.

Both tests were performed on what appears to be the same system, meaning we have two trials here rather than a comparison of how the RTX 3070 runs on different CPUs. The system uses an Intel Core i9-10900K processor, 16GB of DDR4 SDRAM and of course has an RTX 3070 with 8GB of memory for its GPU (though we’re not sure which manufacturer it comes from).

The results? One of the tests has an OpenCL score of 359,349 while the other has an OpenCL score of 350,093. We’re not sure what order these tests were performed in.

 

We first stumbled across these results thanks to Apisak on Twitter. The top (and currently only) reply to Apisak’s thread also took the liberty of answering the burning question on everyone's mind by checking these results against existing benchmarks for the 2080 Ti, which they found normally runs in the 345k-355k range. When we checked ourselves, we found this does appear to be the case on the Geekbench Browser at least, though there are outliers that are both significantly higher and lower than the RTX 3070’s scores.

We should note that it’s difficult to put much weight behind any one single test, as benchmarks vary from system to system based on factors like CPU, cooling and even form factor. But the general trend, both from these new RTX 3070 benchmarks and from the official Nvidia numbers released late last month, is that the RTX 3070 tends to perform slightly better than the RTX 2080 Ti, which would likely make it one of the best graphics cards and put it in a strong position in our GPU benchmarks hierarchy.

And that’s good news, even if the new card’s performance gains aren’t too hefty. That’s because the RTX 3070 is set to start at just $500, which is significantly cheaper than most RTX 2080 Ti listings, which go for more than $1000 even 2 years after its release.

Of course, we still have a small sample size to work with, and official Nvidia numbers are likely to be biased. We can’t say for sure how well the RTX 3070 performs until we publish our official review. But going off preliminary testing, it seems like catching up to today’s top-performers will soon be far more affordable -- assuming inventory catches up to demand.

  • btmedic04
    Its unfortunate that unless you are willing and able to camp out a brick and mortar store a few days prior to launch, their is little likelihood of actually being able to purchase one. Nvidia doesn't care who buys the cards and only said they were manually removing bot purchases to say that that did something. I'd love to replace my GTX 1080 to step into ray tracing in preparation for cyberpunk 2077, but its looking like iI'll be waiting to see what AMD brings to the table with big navi.
    Reply
  • nofanneeded
    No it is not better in anyway .. you cant compare 8GB card against 11GB and say it is better.

    If you can get RTX 2080 ti for $500 it is the better choice.
    Reply
  • thisisaname
    You will be able to buy it before Christmas, just not saying which year that Christmas will be in ;)
    Reply
  • drewthebrave
    This is good news. Hopefully the AMD card reveal next week offers something competitive and hopefully is available to purchase in reasonable quantities so that we can upgrade our rigs for Cyberpunk.
    Reply
  • Homer J.
    nofanneeded said:
    No it is not better in anyway .. you cant compare 8GB card against 11GB and say it is better.

    If you can get RTX 2080 ti for $500 it is the better choice.
    If the 8 GB delivers better fps on 1440p and 4k than the 8 GB card is better than the 11 GB card. Just because something has more RAM does not automatically make it better.
    That is like saying an i7 9700 with 64 GB RAM is better than a Ryzen 5800 with 32 GB RAM. Plus you won´t get a 2080 ti for 500 unless it is used in which case you get the worse choice.
    Reply
  • nofanneeded
    Homer J. said:
    If the 8 GB delivers better fps on 1440p and 4k than the 8 GB card is better than the 11 GB card. Just because something has more RAM does not automatically make it better.
    That is like saying an i7 9700 with 64 GB RAM is better than a Ryzen 5800 with 32 GB RAM. Plus you won´t get a 2080 ti for 500 unless it is used in which case you get the worse choice.

    2% better does not make much difference .,more VRAM is alot better . besides , some games are using more than 8GB already with everything maxed out in 1440p , and for sure future games will need more memory
    Reply
  • VforV
    nofanneeded said:
    No it is not better in anyway .. you cant compare 8GB card against 11GB and say it is better.

    If you can get RTX 2080 ti for $500 it is the better choice.
    Assuming the performance is real and the reviews will tell the same story, actually the RTX 3070 is better than a RTX 2080ti in every way except one: 11GB for 4k gaming.

    In every other way the RTX 3070 is better (for 1080p/1440p gaming):
    you actually need to find a RTX 2080ti at 500$ first, 3070 has 500$ MSRP already.
    you need to play the used market lottery with the 2080ti, 3070 is brand new so the 2080ti can be a dud or will have little to no warranty anymore, the opposite is true with the 3070
    the 3070 will age better and get better driver support as drivers and optimizations get more mature for it
    the 3070 will have higher re-salebility in 2 years, when 2080ti will be 4 years old...
    the 3070 will have lower power consumption than the 2080ti
    Of all these reasons the most important one is that you should always buy new instead of used if the performance is equal.
    Reply
  • sizzling
    nofanneeded said:
    2% better does not make much difference .,more VRAM is alot better . besides , some games are using more than 8GB already with everything maxed out in 1440p , and for sure future games will need more memory
    Which games? On my 3080 I have not seen any go over 6GB using 1440p. Also VRAM used does not mean VRAM required.

    Also Id expect in Ray Tracing the 3070 should be significantly stronger even if non RT games are similar performance.
    Reply
  • colson79
    nofanneeded said:
    No it is not better in anyway .. you cant compare 8GB card against 11GB and say it is better.

    If you can get RTX 2080 ti for $500 it is the better choice.
    Ram doesn't mean anything unless you need it. If the 3070 is as fast as the 2080 ti in normal games and has faster performance in ray tracing it is better. It's been proven that most games don't really need more than 8 GB of vram up to 4k
    Reply
  • colson79
    nofanneeded said:
    2% better does not make much difference .,more VRAM is alot better . besides , some games are using more than 8GB already with everything maxed out in 1440p , and for sure future games will need more memory
    This simple isn't true. Most of the tools that you can get to show VRAM shows you how much VRAM is allocated not how much VRAM is used. If the card has more VRAM that 8 GB it is going to allocate more that 8 GB, but it doesn't mean it's using more that 8GB. You need special development tools to see the actual usages and some test have been done and almost all AAA games use less than 8 GB at 4k ultra.
    Reply