20-core RTX Spark beats mobile chips in Geekbench

Nvidia’s RTX Spark Chip Drops New Geekbench Scores, Signaling A Shift in Laptop Superchip Competition

The race for the next generation of laptop superchips is heating up, and Nvidia has just dropped some intriguing new performance numbers for its upcoming RTX Spark hardware. Recent listings on the Geekbench database have revealed more than one configuration, confirming that the upcoming chip family includes a flexible approach to core design. These preliminary results point toward impressive multi-core performance, though a closer look at the single-core benchmarks reveals a fascinating story about where future AI silicon will truly compete.

Nvidia has presented two distinct SKUs in the testing phase. The first option features a 20-core configuration, boasting clocks of 4.0 GHz and scoring 2,570 points in the single-core test and 23,126 points in the multi-core assessment. The second variation is an 18-core model, utilizing 10+8 core clusters clocked at 3.9 GHz, achieving a multi-core score of 21,776. While both variants show excellent multi-core capacity, their single-core performance sits remarkably close together.

When stacked against other Windows-running System on a Chip (SoC) contenders, these Nvidia scores place the RTX Spark squarely among the elite. Although systems like the Snapdragon X2 Elite Extreme currently hold the top spot with 25,075 points, even the Nvidia variants demonstrate significant power. For instance, when measured against the AMD AI Max+ 395 chip, the 18-core Nvidia SKU is still 5.6% faster in multi-core processing, and the 20-core variant edges it out by about 12%.

The performance context deepens when comparing these results to established high-end rivals. The 20-core RTX Spark holds a 3.3% advantage over the 14-core M3 Max from Apple, a machine that scores 22,385 points in the multi-core test. This suggests that even with unoptimized drivers, Nvidia is positioning its upcoming silicon to compete fiercely against established high-performance competitors.

The implications for potential future hardware are substantial. Speculation suggests that as driver optimization fully takes hold upon release, the RTX Spark could eventually surpass current Apple Silicon offerings. Indeed, the 18-core SKU’s multi-core score of 21,776 is equal to, or potentially greater than, the base M4 Pro, positioning it strongly against current generation Mac chips.

On the single-core front, the landscape shifts dramatically. While Nvidia’s scores are less dominant compared to Apple’s latest silicon—where the base M5 chip achieves a remarkable 41% lead over the RTX Spark test score—the performance gap narrows when looking at other architectures. The Ryzen AI Max+ 395 only scores 2,429 points in single-core testing, making the Nvidia chips notably faster by 5.8% and 4.6%, respectively.

This indicates that while raw single-core speed might not be the immediate focus, the multi-core architecture and potential for powerful AI integration are where the RTX Spark is making its strongest claim. The consistent improvement across these preliminary benchmarks suggests that Nvidia is laying a solid foundation for a superchip that will deliver cutting-edge performance when it officially launches, especially with promises centered on integrating DRM and anti-cheat technologies natively on Windows.