Intel Panther Lake Core Ultra X9 388H performance analysis - Outpaces Arrow Lake and exceeds Zen 5 in efficiency

The new Panther Lake mobile processors are once again made up of different tiles, with the compute tile being produced by Intel itself (Intel 18A). Depending on the model, up to 16 cores in total can be used: four fast P-cores (Cougar Cove), eight E-cores (Darkmont), and four additional low-power E-cores (Darkmont). NPUs with 46-50 TOPS are also installed, meaning that all Panther Lake processors meet the minimum requirement for Microsoft's Copilot+ certification. TSMC manufactures the other two tiles (the platform controller and the GPU).

There are differences between the various models when it comes to RAM. The Core Ultra X chips with the fast GPUs are only available with LPDDR5x RAM (8533 or 9600 MT/s) and a maximum of 96 GB. The other models with the small iGPU, on the other hand, can be combined with either LPDDR5x RAM (6800 to 8533 MT/s) or SO-DIMM modules (6400 to 7200 MT/s), and support up to 128 GB of RAM.

The new Panther Lake chips are designed for a TDP range of 25 to a maximum of 80 watts. This is an advantage in our eyes, as the extremely high power limits of the Arrow Lake chips (up to 115 watts) are now a thing of the past.

The new CPUs are available with three different integrated graphics cards. The fast models Arc B390 (12 Xe cores) and Arc B370 (10 Xe cores) are limited to a few models; all other CPUs only use a GPU with 4 Xe cores. So you have to be careful when buying a new laptop. A more detailed look at the new iGPUs is available in a separate article.

As a test system with the Core Ultra X9 388H, we had the Asus Zenbook Duo UX8407 at our disposal, which we have already extensively tested earlier. Here, the processor is paired with 32 GB of fast LPDDR5x-9600 RAM, and the power limits are set to 64/46 watts. There will therefore be even faster implementations.

In order to make a meaningful comparison between the different processors, we consider both the power consumption and the actual performance in synthetic benchmarks, from which we then determine the efficiency. Consumption measurements are always carried out on an external display to eliminate differences among internal displays as influencing factors. Nevertheless, we measure the system's overall consumption here and do not just compare the pure TDP values.

We start with the single-core performance, and here the new Panther Lake processor is only just ahead of its predecessors. This makes it currently the fastest mobile processor in this TDP range, apart from Apple's M4 and M5 generation, which are in a different league. The fastest Snapdragon X processor, the X1E-84-100, is also narrowly beaten, and the new AMD Ryzen AI 9 465 is around 8 percent slower.

To achieve such a performance level, the X9 388H requires around 11-12 watts, which corresponds to a package power of 13-15 watts. This means the Panther Lake chip is more efficient in single-core scenarios than its Arrow Lake predecessors and AMD's Zen 5 chips. However, the Lunar Lake CPUs still have a slight advantage, just as the entire ARM competition from Qualcomm and Apple.

Performance has also increased in the multi-core tests, and only the very fast Arrow Lake chips and Zen 5 CPUs with high power limits can keep up with the new X9 388H, although the power limits here are significantly lower at 64/46 watts. The Lunar Lake CPUs are completely left behind in terms of pure performance, and the X9 388H is practically twice as fast here. The performance is also on par with the Apple M5 SoC, which, however, operates at significantly lower power limits (max. 30 watts).

When it comes to multi-core efficiency, you have to look at the TDP values as well as the pure performance. For the X9 388H, in addition to the standard power limits, we also conducted measurements with fixed power limits of 20, 28, 35, and 45 watts. We also carried out additional measurements for the new Ryzen AI 9 465 at 20, 28 and 35 watts. Of course, the efficiency increases with lower power limits.

The efficiency of the new X9 388H has improved overall. If you look at the power limits, for example, you can see that the X9 is more efficient at 45 watts than the Ryzen AI 9 465 at 35 watts, and also slightly better at 35 watts than the HX370 at 33/28 watts. At very low power limits, even the Apple M4's efficiency is almost reached.

Let's take a closer look at the performance at lower power limits. Here you can clearly see that the new X9 388H has a performance advantage over AMD's Zen 5 models even at low power limits, a trend that was not previously observed. However, the difference to the Ryzen AI 9 HX370 is small, so Intel has only caught up here. Nevertheless, we can clearly say that Panther Lake can replace its two predecessors (Arrow Lake and Lunar Lake) very well and is also a very good choice in slim devices with low power limits. Together with the excellent GPU performance, Panther Lake will clearly be the best platform for handhelds in 2026.

With the new Panther Lake mobile processors, Intel has managed to successfully combine the two previous generations, Arrow Lake and Lunar Lake, as the performance is even better than with Arrow Lake, while efficiency has been improved at the same time. Even with low power limits, the performance is very competitive, and Intel (in conjunction with the new GPUs) is therefore the better choice for slim laptops.

AMD's Zen 5 processors fall slightly behind in single-core performance, but AMD is still competitive in multi-core performance. Apple's current M5 SoC is at the same level in multi-core performance but is more efficient, while single-core performance is in a completely different league. The upcoming M5 Pro SoCs will certainly raise the bar for multi-core and GPU performance.

We will have to wait a little longer for the new Snapdragon processors from ARM, but these will probably have a performance advantage over Panther Lake. However, devices with Panther Lake are already coming onto the market and are currently the best choice when buying a new laptop.