ASUS ProArt P16 RTX Spark review

For content creators, the additional resolution also gives you more room to work with detailed images and 4K video footage, which can be useful when editing. That said, I still think the 3K OLED looks excellent in its own right, and unless you’re particularly demanding about display quality, you probably won’t feel like you’re missing much.

There is also the question of battery life. A higher-resolution display has more pixels to drive, although that does not automatically mean it will always consume more power. Display brightness, content and panel efficiency also play a part, which is why I wanted to test both models rather than assume that the 3K version would last longer.

Unfortunately, the 3K OLED configuration is not officially available in Singapore, at least for now. That is a shame because it gives buyers another option, particularly those who want RTX Spark’s processing capabilities but do not necessarily need the more expensive Tandem OLED panel.

Performance and battery life

As I mentioned in my earlier RTX Spark impressions, the ProArt P16 is not quite like a conventional Windows laptop. The RTX Spark N1X processor in our review configuration has a 20-core Arm CPU and a Blackwell GPU with 6,144 CUDA cores. The version I previously tested also came with 64GB of unified memory, which allows the CPU and GPU to share a common pool rather than rely on a separate allocation of dedicated graphics memory.

That makes a difference when working with large AI models, but it is also worth remembering that this is still Windows on Arm. While Microsoft’s compatibility layer has improved considerably, some applications continue to run through emulation, while others may not work properly at all.

Fortunately, most everyday applications are not particularly demanding for this hardware. Browsing with multiple tabs open, working across Microsoft Office applications and editing photographs should pose little to no difficulty. What interested me more was how the P16 would perform against Apple’s 16-inch MacBook Pro with the M5 Max, which offers a similar combination of high-performance processing and unified memory.

For productivity testing, I have included UL Procyon’s Office Productivity benchmark. It uses actual Microsoft Word, Excel, PowerPoint and Outlook workloads, making it a more useful measure of everyday responsiveness than relying entirely on synthetic CPU scores.

Interestingly, the 4K Tandem OLED model score just a tad higher than the 3K OLED version – just shy of 7%. That seems rather unusual considering that both laptops have essentially identical hardware apart from their displays. I wouldn’t read too much into this result, however, since a higher-resolution screen shouldn’t automatically make a laptop faster at office tasks. But suffice to say, those are pretty good scores.

Gaming is another area where RTX Spark has a lot to offer. During my earlier testing, I was able to run Cyberpunk 2077 at around 75fps at 1440p with medium settings, which is pretty good for a laptop that does not rely on a discrete graphics card.

For a new test against the MacBook Pro with M5 Max, I upped the ante running the game at higher details – albeit at a lower 1080p resolution. As you can see in the chart above, the Apple machine ran smokes around the RTX Spark laptop, although it is important to remember that the Windows version of the game is running through emulation, while the same game runs natively on MacOS. But that does not mean the P16 is a poor gaming laptop. In fact, having access to NVIDIA’s DLSS and ray-tracing technologies makes it a pretty capable machine for playing games. However, if you’re looking to buy a laptop for gaming, then there are better options out there that also cost far less.

But perhaps the more interesting comparison is how both laptops perform when running AI models locally. After all, that is one of the key selling points of the RTX Spark in the first place, and having access to CUDA and up to 128GB of unified memory makes the ProArt P16 an interesting alternative to Apple’s MacBook Pro.

To see how the two platforms compare, I ran llama.cpp on both laptops to see how quickly a model processes the information on them, and how quickly it generates a response. Both are measured in tokens per second, but they tell us different things about AI performance.

Interestingly, the results were mixed. Using a smaller 9-billion-parameter model, the ProArt P16 processed prompts at higher tokens per second, compared to the MacBook Pro.

However, the situation reverses when the AI starts generating its response. The MacBook Pro delivered almost double the tokens per second against the ProArt, making it substantially faster at producing text. The difference becomes even more interesting with a larger 27-billion-parameter model, where the M5 Max also produced more tokens per second than the RTX Spark.

But I don’t think these numbers necessarily tell us which laptop is better for local AI. In the same tests, the ProArt was able to begin answering a prompt containing a 50-page document faster than the MacBook Pro. By the time both machines completed their responses, however, they had taken roughly the same amount of time.

What this tells me is that RTX Spark’s strength isn’t necessarily in generating text faster than Apple’s silicon. Rather, its ability to process large prompts quickly, combined with NVIDIA’s CUDA ecosystem and unified memory architecture, makes it particularly useful for certain AI workloads. Apple’s M5 Max is still a formidable competitor, especially when it comes to generating responses, so I wouldn’t say RTX Spark has a clear performance advantage across the board.

And that is worth remembering if you are considering the ProArt P16 primarily for AI development. Having a powerful GPU and plenty of memory is certainly useful, but the applications and models you intend to run will ultimately determine whether NVIDIA’s platform makes more sense than a MacBook Pro.

Going back to the Battery life is another area where I wanted to see how the two displays affect the overall experience. Both P16 variants use a large 99Wh battery, but the difference in display technology could make one more efficient than the other.

And I have to say the difference in the Procyon’s Office Productivity Battery Life test was quite substantial. It makes the 3K model a more interesting proposition than I initially thought. The 4K Tandem OLED certainly has its advantages when it comes to resolution and brightness, but you’re also giving up a considerable amount of battery life when using the display at its maximum setting.

Final thoughts

There are two ways I look at the ASUS ProArt P16. The first is simply as a laptop, and from that perspective, I think ASUS once again has done a great job. It looks professional, is relatively thin and light for a 16-inch machine, has a useful selection of ports and offers a large OLED display that is well suited to creative work.

The second is as one of the first laptops powered by NVIDIA’s RTX Spark, and that is where things become a little more complicated.

I like what NVIDIA is trying to achieve here. Bringing the Grace CPU, Blackwell GPU and unified memory together truly elevates Windows laptops for developers and creators who want to work with AI locally. It is not merely about running a chatbot on your laptop, but having the hardware and software ecosystem to experiment with larger models without relying entirely on cloud services.

At the same time, I do not think RTX Spark is something everybody needs. If your work mainly involves Microsoft Office, web browsing, photo editing and occasional video work, there are plenty of less expensive Intel- and AMD-powered laptops that will do the job just as well. And while the P16 can handle games, it is not necessarily a better buy than, say, an ROG gaming latop.

The comparison with Apple’s M5 Max MacBook Pro is also revealing. Apple’s processor has a clear advantage in several conventional CPU workloads, while the P16’s appeal lies more in NVIDIA’s GPU capabilities, CUDA support and flexibility for local AI development. Neither platform is automatically the better choice; it depends on which applications you actually use.

Then there is the question of the displays. I do not think the 4K Tandem OLED is necessarily the one everybody should buy simply because it is technically superior. The 3K OLED already offers a high resolution, wide colour coverage and a 120Hz refresh rate, and for many people, that should be more than sufficient.

And having seen the battery life results, I think the 3K version actually makes a stronger case for itself. Almost two additional hours of battery life under our demanding test conditions is not something I would dismiss lightly, especially when the difference in resolution may not be particularly noticeable to many users during everyday work. It’s unfortunate that ASUS isn’t bringing this configuration to Singapore, because I suspect it would have appealed to buyers who value battery life over having the higher-resolution display – and cost lesser too.

In Singapore, the 64GB version of the ProArt P16 starts at $7,499, while the range starts at $6,599 for the 48GB configuration and rises to $11,099 for the 128GB model. Those are considerable amounts of money to spend on a laptop, especially when you can buy a capable Windows workstation or MacBook Pro for less.

But the ProArt P16 is not really trying to appeal to everybody. It is a premium creator laptop that also serves as a showcase for what NVIDIA believes the next generation of Windows on Arm computers can do. And while I still think RTX Spark has some convincing to do, particularly when it comes to value, the P16 makes a pretty good case for the platform.

The ASUS ProArt P16 (H7607) is available at the ASUS webstore and authorised dealers now. A 14-inch version of the laptop, the ProArt 14, is available now as well.