NVIDIA RTX Spark N1X Outpaces Ryzen AI Max+ 395 and Panther Lake X9 388H in Leaked Geekbench 7 Multi-Core

RTX Spark N1X Appears in Two Configurations

Two NVIDIA RTX Spark N1X systems have appeared in Geekbench 7 running Windows 11 on Arm. The faster system uses the full 20-core CPU configuration and a 6,144-core Blackwell GPU. It scores 2,570 points in single-core and 23,126 points in multi-core.

NVIDIA RTX Spark N1X Outpaces Ryzen AI Max+ 395 and Panther Lake X9 388H in Leaked Geekbench 7 Multi-Core
NVIDIA RTX Spark N1X Outpaces Ryzen AI Max+ 395 and Panther Lake X9 388H in Leaked Geekbench 7 Multi-Core

The second result shows an 18-core CPU paired with a 5,120-core GPU. This version scores 2,541 and 21,776 points respectively. Geekbench reports a 10+10 CPU cluster configuration for the full chip and 10+8 for the cut-down version. Both systems have 64GB of memory, while their listed base frequencies are 4.0 GHz and 3.9 GHz.


RTX Spark Ahead of Ryzen AI Max+ 395 in Geekbench 7

The Ryzen AI Max+ 395 is one of the closest existing Windows comparisons. It combines 16 Zen 5 cores with Radeon 8060S graphics featuring 40 RDNA 3.5 compute units and supports up to 128GB of LPDDR5X memory. N1X is about 6% faster single-core and 12% faster multi-core than the current Geekbench average.

Intel’s flagship Panther Lake mobile processor takes a different approach but also combines a high-end CPU with large integrated graphics. The Core Ultra X9 388H has 16 CPU cores in a 4P+8E+4LPE configuration and an Arc B390 GPU. Its available Geekbench 7 result scores 2,767 single-core and 19,304 multi-core. N1X is therefore around 7% slower in single-core but 20% faster in multi-core, although Intel’s number currently comes from a single public result rather than an established average.


Apple M5 Max Remains Far Ahead

Apple’s M5 Max provides another useful comparison because it combines an 18-core CPU, up to a 40-core GPU and as much as 128GB of unified memory in one package. Across 19 public Geekbench 7 M5 Max submissions checked, the chip averages approximately 3,766 single-core and 35,527 multi-core. This leaves N1X around 32% behind in single-core and 35% behind in multi-core.

Also, Read


What These Scores Mean

The 20-core N1X comes within 11% of the 14-core M4 Max in multi-core performance, a significant improvement over earlier tests. However, the single-core gap to Apple’s chips remains substantial at around 24.5%.

It is worth noting that Geekbench primarily evaluates CPU performance, while the N1X’s Blackwell GPU with 6,144 CUDA cores could provide a serious advantage in AI and graphics workloads that these scores do not reflect. Additionally, these tests were conducted on prototype systems with 64GB of unified memory running in balanced power mode, and drivers may still be unoptimized.

At this stage, these scores position the N1X as an Arm Windows chip capable of competing with the best mobile x86 offerings in multi-threaded compute, while integrating a massive graphics block. If these configurations reach finished products with manageable thermal envelopes, NVIDIA could be aiming for more than just a software demonstrator for Windows on Arm.

Source: Geekbench

Leave a Comment