Intel Core Ultra 9 285K processor review: Good power efficiency, mixed performance

Note: This review was first published on 24 October 2024.

Finally, Intel’s highly anticipated next-generation processors are here. The Core Ultra 200S, codenamed Arrow Lake, is built on the new 4nm process by TSMC and uses a new tile-base design instead of the monolithic approach seen with Raptor Lake. So the performance cores (P-core) are based on a new architecture named Lion Cove, and the efficiency cores (E-core) adopt the Skymont design. If both names sound familiar, that’s because this is essentially the same architecture as the recent Lunar Lake notebook processors.

I’ve already covered the Core Ultra 200S extensively in a separate feature which you can read here (and I highly recommend you do so first), so I’ll jump straight to what we are here for: the flagship Core Ultra 9 285K’s benchmarks and numbers!

But first, a quick price comparison between the new and the old. Intel has tagged the Core Ultra 9 285K at the same launch price as its predecessor, the Core i9-14900K, at US$589. It’s certainly an interesting strategy that raises eyebrows, when you consider that its competitor AMD launched their Ryzen 9000 series at lower price points than their last-gen 7000 series, and works with existing socket AM5 – so it’s not necessary to buy a new X870 motherboard unless you want the latest bells and whistles like Wi-Fi 7.

The Core Ultra 200S chips, on the other hand, only works with the new Z890-based motherboards like the ROG Maximus Z890 Hero (read about it here). So you’re looking at quite a hefty investment if you’re considering to upgrade from your existing Intel setup, and maybe even less compelling if your AMD setup is no more than two years old. But from a performance perspective, there are many things to admire about what Intel has done with Arrow Lake, especially when taking into account how poor its Raptor Lake generations were.

Let’s take a look.

Test Rig

Our CPU test rig comes with the following specifications, with motherboards used being the only different components used.

• ROG Crosshair X870E Hero (AMD) / ROG Maximus Z890 Hero motherboard (Intel)
• Samsung 980 Pro 1TB SSD
• Kingston Fury 32GB DDR5 memory
• NVIDIA GeForce RTX 4090 Founders Edition
• Windows 11 OS

For cooling, we used the ROG Strix LC III ARGB LCD for both Intel and AMD processors. I then put the Core Ultra 9 285K through some of our regular gaming and workload benchmarks and against AMD’s bests – the Ryzen 9 9900X and 9950X.

Gaming performance

My game list includes a mixture of old and recent games to test these processors’ performance, and while the list isn’t exhaustive by any measure there are enough different game engines and APIs variety to give us an idea of broader performance trends.

1080p benchmarking is a great measure of a CPU’s prowess, no thanks in part to the high-performance GeForce RTX 4090 card used here. You see, at lower resolutions, the GPU can process and transfer data much quicker than at higher resolutions. A CPU bottleneck happens here because the processor cannot keep up with the processing speed of the graphics card. The CPU, after all, is responsible for processing real-time game actions, physics, UI, audio and other complex CPU-bound processes.

So while it’s unlikely that Core Ultra 9 285K owners are going to be playing games at 1080p, the results at this resolution can clue us in the chip’s the raw power when it comes to gaming.

The two Ryzen 9s pulled ahead of both Intel processors here.

Total War: Warhammer III is the only game where we experienced a weird issue where the game struggles to run at 4K with the Core Ultra 9. Crawling at 33fps is highly unusual and not a true reflection of the CPU and GPU capabilities. This issue did not show up with the Ryzen 9 chips and the i9-14900K.

That said, the Core Ultra 9 285K performed the best here.

The Core Ultra 9 285K could almost match the Ryzen 9 9900X but not the 9950X.

It’s the same with Returnal.

Another rare win for the Core Ultra 9 285K, but it was the i9-14900K that surprised me here.

The Core Ultra 9 285K seems to really struggle with Cyberpunk 2077.

Productivity and Content Creation performance

While the Ryzen 9 9900X and 9950X are still the better processors when it comes to gaming, it’s quite a different story in our real-world productivity and content creation benchmarks.

In Handbrake where each processor encodes a 4.5GB video, the Core Ultra 9 285K came out the fastest – besting the Ryzen 9 9950X. Despite having lesser cores here, it does look like the combined E and P cores work more efficiently for this kind of workload.

SYSmark 30 is a great benchmark as it measures and compares system performance using real-world applications and workloads – such as office suite-style applications, tasks like web browsing, file compression, and application installation, and photo and video editing applications, including multitasking.

Here, the results are almost identical to what we saw with Handbrake. The Core Ultra 9 285K performs really well with content creation workloads, such as video editing. But it lags behind the Ryzen 9 9950X processor – though not that far off from the 9900X – when it comes to other simpler tasks, such as file compressions and even simple photo editing.

Power and thermal efficiency

I’ve already talked about how Intel has built the Core Ultra 200S series focusing on power efficiency with a twist: you won’t be getting peak performance and peak efficiency simultaneously (read about it here). But just comparing gen-to-gen, the Core Ultra 9 285K is a markedly improvement over the i9-14900K in both power and thermal efficiencies.

As we can see in the charts above, where I measured the power and temperature of each processor while running Cinebench R23 (Multi Core) and Metro Exodus, all four processors exceeded its TDP and the 9900X and i9-14900K even hit their PPT (Package Power Tracking). Not the 9950X (230W) and certainly not the Core Ultra 9 285K, with the latter’s PPT limit being 250W. There’s certainly room for overclocking potential with the Ultra 9, unlike the Raptor Lake processors.

All in all, I think Intel deserves some plaudits for finally making a more power efficient processor that also has headroom for overclocking.

Final thoughts

So, there you have it – a confounding mix of gaming results that is rather puzzling. A couple of titles fared well on the Core Ultra 9 285K while others performed poorly. It’s clear that Intel has got some issues here to iron out (I mean, the Total War: Warhammer III problem at 4K can’t just be a one-off). As it stands, the chip’s gaming performance is inconsistent at best, but productivity and content creation tasks are where it stands out – especially in the latter category.

And then there’s the matter of power efficiency, where the Core Ultra 9 285K improves spectacularly over Raptor Lake. This is great to see, although there’s still plenty of ground for Intel to cover if they want to match AMD on this front.

That’s not to say it’s all doom and gloom, however. Now that Intel has shifted gears with its CPU designs and is moving away from custom in-house process nodes to working with TSCM, some growing pains are to be expected. While the switch to a new external foundry might explain some of the slight dips in clock speeds and overall performance – even if it yielded notable power efficiencies – there’s no denying that chiplet-based CPUs are the way forward – as AMD has shown with Ryzen. Any early hiccups with Arrow Lake should ultimately pave the way for more refined and compelling iterations down the road.

Intel’s immediate concern should be the price. At US$589, it's harder to recommend the Core Ultra 9 285K than I would have liked, when performance is mixed at best when compared to AMD's Ryzen 9 9900X and 9950X. More so now that AMD has decided to spoil the party with a price cut for both chips.

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.