Asus ProArt PX13 review

Frame rates

1920×1080, High preset, upscaling off. Averages 106fps across the 5 core benchmark games.

Measured on this machine. Source: Notebookcheck, 2024-07-29.

In use

The short version

The Asus ProArt PX13 is not a gaming laptop that happens to be small. It is a 13.3in creator machine with a serious graphics card fitted, and almost every judgement worth making about it follows from that ordering. The specification reads well in isolation: an RTX 4070, a Ryzen AI 9 HX 370, 32GB of memory, a 2TB SSD, and a 2880x1800 OLED panel, all inside a chassis weighing 1.389kg. But the panel runs at 60Hz, and that single figure reorganises everything else.

Notebookcheck published ten games for this machine on 29 July 2024, at 1920x1080 on the High preset with upscaling switched off. The range runs from 217fps in Strange Brigade down to 62fps in Horizon Forbidden West. Unusually for our database, the full five-game scoring suite is present, and it averages 106fps. That is a complete, properly comparable set of measurements.

The most interesting thing about that data is not the top of it. It is that every single measured result clears the panel's 60Hz refresh rate, including the worst one. On its own screen, at the resolution these figures were recorded at, this machine never runs out of frames. That is a genuinely unusual thing to be able to say, and it is the strongest argument for the PX13 as a games machine.

The counter-argument is equally clear-cut. The screen is 2880x1800, not 1920x1080, and that is a very large gap. It is also 60Hz, which means the enormous headroom in the light half of the chart — 217fps, 201fps, 167fps — delivers nothing at all on the built-in display. And the comparison data suggests this RTX 4070 is not running anywhere near the power limits found in larger machines carrying the same card. This is a portable computer with capable graphics, not a fast gaming laptop in a small box.

One point of housekeeping: we have no confirmed UK price for this exact configuration, so we cannot publish a score for it. Value is a substantial part of how we score machines and it cannot be computed without a price. The benchmark suite is complete and the laptop ranks normally on frame rate; the missing element is the price, not the data.

Performance

The power limit we do not have

Our specification for this machine does not include a total graphics power figure, and that is a meaningful gap rather than a formality. On this generation of mobile hardware, the power limit is the single most predictive number in the specification — more predictive than the model name, which was applied to implementations drawing wildly different amounts of power. Without it, the honest position is that we know which graphics processor is fitted and we do not know how hard it is being allowed to work.

What we can do is read backwards from the measurements, and the measurements are informative. As the comparison section below sets out in detail, this machine lands within a frame of an RTX 4060 machine running at 85W across five shared games — and those five games include the two heaviest titles in the suite. A tie over that set is not the kind of tie you can dismiss as a quirk of light workloads. It is consistent with a graphics card operating well below the ceilings that larger chassis allow.

That is not a scandal, and it is not hidden. A 13.3in laptop weighing 1.389kg has a hard physical limit on how much heat it can move, and the graphics card has to share that budget with a modern eight-core-class processor. Something has to give, and in a machine this size it is always the sustained graphics power. What matters for a buyer is understanding that you are buying an RTX 4070 in the same sense that you buy an engine in a small car: the part is capable of more than the vehicle around it will let it deliver.

Reading the chart against the right ceiling

Ten measured games, at 1920x1080 on High with no upscaling: Strange Brigade 217fps, The Witcher 3 201fps, GTA V 167fps, Dota 2 132fps, Far Cry 5 119fps, Final Fantasy XV 104fps, Cyberpunk 2077 86fps, X-Plane 11 74fps, Baldur's Gate 3 73fps, Horizon Forbidden West 62fps.

On most gaming laptops the useful exercise is to compare those figures against a high refresh rate and ask which titles clear it. Here the exercise is almost trivial, because the refresh rate is 60Hz and the lowest measured result is 62fps. Every title in the suite is above the panel. The machine is, on its own display, never short of frames.

It follows that the top three-quarters of this chart is doing nothing for you on the built-in screen. Strange Brigade at 217fps and The Witcher 3 at 201fps represent frames the panel will never show. That is not wasted in the abstract — headroom is what stops a game dipping below refresh during heavy scenes, and headroom of that scale means those two titles will simply never dip — but it is not performance you experience. Anyone choosing this machine on the strength of its higher numbers is choosing on the basis of something the hardware in front of them cannot express.

Where the headroom does become real is on an external display. A machine like this is very likely to spend a large part of its life connected to a monitor, and a high-refresh external panel would turn the top half of that chart into something you can actually see. That is a legitimate use case and it changes the calculation considerably. But it is a use case that involves carrying the laptop to a desk and plugging it in, which is worth being clear about when the machine's entire proposition is that it weighs 1.389kg.

The resolution problem

The larger caveat is resolution. Every figure above was recorded at 1920x1080. The panel is 2880x1800, which carries roughly one hundred and thirty per cent more pixels — more than double the work per frame in any title that is genuinely graphics-limited.

For the light half of the chart this barely matters. Strange Brigade, The Witcher 3 and GTA V have so much margin over 60Hz that even a severe reduction leaves them comfortably above the panel. For the heavy half it matters enormously. Cyberpunk 2077 at 86fps, X-Plane 11 at 74fps, Baldur's Gate 3 at 73fps and Horizon Forbidden West at 62fps do not have the margin to absorb a doubling of pixel count and stay above 60fps. At native resolution, with these settings and no upscaling, those four titles will be below the refresh rate.

The remedy is the obvious one and it is entirely reasonable: run games at a lower rendering resolution, or leave upscaling switched on, and let the panel do what a high-density panel is good at. On a 13.3in screen the pixel density is high enough that a rendered resolution below native is far less visually obvious than it would be on a 17in display. But it does mean that the honest description of this machine's gaming behaviour is "runs everything measured above its refresh rate at 1080p", not "runs everything at native resolution", and those are different claims.

Where the spread comes from

The distance from 217fps to 62fps is a factor of three and a half on unchanged hardware at unchanged settings, and the reasons are worth separating because they are not all the same reason.

The first group is titles that are not really testing the graphics card. GTA V at 167fps and Dota 2 at 132fps sit high because they are old or deliberately lightweight, and at 1920x1080 the graphics card finishes each frame and waits for the processor to supply the next. Dota 2 at 132fps is actually a modest result in absolute terms for a game of its weight, which tells you something about the platform rather than about the GPU — competitive titles of that kind are limited by single-thread throughput and driver overhead more than by shader performance.

The second group is titles that are testing it and finding it adequate. Far Cry 5 at 119fps and Final Fantasy XV at 104fps are genuine graphics workloads from a few years back, and comfortably handled.

The third group is where the machine's real limits appear. Cyberpunk 2077 at 86fps, Baldur's Gate 3 at 73fps and Horizon Forbidden West at 62fps are modern, expensive titles asking for everything available, and the results narrow sharply. The gap between Far Cry 5 at 119fps and Horizon Forbidden West at 62fps — less than a factor of two — is the part of the chart that predicts the future, because future games look like the second of those, not the first.

X-Plane 11 at 74fps sits oddly in that grouping and deserves its own note. A flight simulator of that vintage is dominated by single-thread processor performance and by how fast scenery can be streamed and prepared, not by shader throughput. Its low position here is not evidence that the graphics card is struggling; it is evidence that this class of software has never scaled with graphics hardware in the way that game engines do.

Baldur's Gate 3 and the complete suite

It is worth flagging that this machine has a figure for Baldur's Gate 3 — 73fps — because a great many laptops in our database do not, and its absence is the most common reason we cannot rank a machine. Having the complete five-title set means the PX13's 106fps scoring average is directly comparable with every other complete entry we hold, on exactly the same basis.

That is a meaningful advantage when you are shopping. A number that has been computed the same way for every machine is worth far more than a headline figure from a press release, and it is the reason we insist on the full set rather than accepting partial averages. The only thing standing between this laptop and a full score is the absence of a confirmed UK price, which is a gap on our side rather than a defect on the machine's.

The screen

A 13.3in OLED at 2880x1800 is a superb panel specification in almost every respect that does not involve refresh rate.

Pixel density at that combination is very high — substantially higher than the 1600p 16in panels found on mainstream gaming machines — which means text rendering, fine detail in photographs and the edges of vector artwork are all rendered without the softness that lower-density displays impose. For the work this machine is named after, that is the point. OLED adds per-pixel illumination, which is what produces genuinely black blacks rather than dark grey ones, and the contrast that follows from that is the single most visible difference between panel technologies in ordinary use.

We do not hold measured panel data for this machine — brightness, colour gamut coverage, colour accuracy and response time are all absent from our figures — so we are not going to make claims about how well it is calibrated. An OLED at this specification in a machine positioned for creative work sets a reasonable expectation, but expectation is not measurement.

The refresh rate is the compromise, and it is a real one. 60Hz is the baseline, and on a machine whose graphics card can produce 217fps in a light title, it is a conspicuous one. The consequence is not that games run badly — as established, every measured title clears 60fps — but that the fluidity advantage of a high-refresh display is simply unavailable. Mouse movement, camera pans and fast motion will look exactly as they do on any ordinary laptop. If you have used a high-refresh display and found the difference significant, you will notice its absence here every day, in the desktop as much as in games.

There is a defensible logic to the choice. High refresh rates cost battery life, and a 13.3in machine with a 73Wh battery has less to spend than a large gaming laptop with a bigger pack. Creative work rarely benefits from refresh rates above 60Hz, whereas it benefits enormously from resolution, contrast and colour. Asus has optimised the panel for the work rather than the play, which is consistent with everything else about the machine. It is the right decision for the buyer this laptop is aimed at, and the wrong one for a buyer who is primarily gaming.

The other consequence of a 60Hz ceiling is that the machine's substantial frame rate surplus in most titles has only one useful outlet: an external monitor. If you already own a high-refresh display, the PX13 will feed it well in the lighter half of this chart. If you do not, a large part of what you are paying for in the graphics card is inaccessible.

Build, size and portability

At 1.389kg, the PX13 is in a different category from almost everything else that carries this class of graphics card, and that is the machine's genuine and substantial selling point.

The practical difference between a 1.389kg laptop and a 2.3kg one is not a small percentage; it is the difference between a computer you carry without thinking and one you plan around. A machine this weight goes in a normal bag alongside everything else you were already carrying. It works on a train table. It does not make you reconsider a day out. Gaming laptops of comparable graphics capability generally do all three of those things badly, and no amount of extra frame rate compensates for a machine you leave at home.

The 13.3in footprint reinforces that. It is small enough to work on an aircraft tray table and small enough to use on a lap without the front edge overhanging. Against that, 13.3in is genuinely small for extended work, and anyone who spends their day in multiple windows will want an external display — which, given the panel's 60Hz limit, they may well want anyway.

The 73Wh battery is modest by gaming laptop standards but sensible for the size, and the combination of a small high-density OLED panel and a mobile processor is a much better basis for real battery life than a large high-refresh display attached to a high-power desktop-class chip. We have no measured runtime figures for this machine, so we will not quote hours in any workload. What we can say from the specification is that this is one of the few laptops in this performance class where unplugged use for non-gaming work is a realistic daily proposition rather than an emergency measure.

Gaming on battery is a different matter, and the honest expectation is that it will be substantially slower than the figures above, as it is on every laptop of this kind. Graphics cards are not permitted to draw full power from a battery, and a 73Wh pack has less to give than the 90Wh and 99Wh packs found in larger machines.

The 2TB SSD and 32GB of memory deserve credit. Both are generous, and both are specifications that small machines frequently skimp on because there is less room for them. 32GB in particular changes what the laptop is good for — video editing, large photo libraries, virtual machines and heavy browser use all stop being constrained by memory, which matters a great deal more for the professional work this machine is aimed at than an extra ten frames a second would.

How it compares

Four machines from the same graphics generation provide the spread. Each comparison uses only games both machines have recorded, and the number of shared titles varies — which, as will become clear, matters more here than the margins do.

Against the MSI Thin GF63 (RTX 4050)

Across five shared games the ProArt PX13 averages 30fps more than the MSI Thin GF63. Five shared titles is a decent sample, and it will include the heavier end of the suite where genuine graphics differences show up, so this is a result you can take at face value.

A 30-frame average advantage over an RTX 4050 machine is a real and useful margin. It is also, in the context of a card two tiers higher, a modest one. An RTX 4070 in a large chassis at a high power limit would be expected to open a considerably wider gap than that over an entry-level part. The size of this margin is the first piece of evidence that the PX13's graphics card is working within tight constraints, and the next comparison makes the case more strongly.

Against the Acer Nitro V 16 (RTX 4060 at 85W)

Across five shared games these two machines are separated by a single frame per second, in the PX13's favour. This is the most important comparison in the set and it deserves careful reading.

Five shared games, in a suite where the scoring titles include Cyberpunk 2077 and Baldur's Gate 3, means this is not a comparison built only on lightweight, processor-limited software. It includes titles that genuinely stretch a graphics card. Under those conditions, a tie is not an artefact. It is a result.

What it says is that an RTX 4070 in a 1.389kg 13.3in chassis performs like an RTX 4060 running at 85W in a 16in one. That is not a criticism of the engineering — extracting RTX 4060-at-85W performance from a machine that weighs what this one weighs is a genuine achievement, and the Acer is a considerably larger and heavier computer. But it should demolish any assumption that the model number on the graphics card tells you what you are getting. If you were choosing between these two purely on graphics performance, there is nothing to choose. You would be deciding on size, weight, screen and price instead, which is exactly the right basis.

It also explains the modest margin over the RTX 4050 machine above. These three results are internally consistent: a constrained RTX 4070 sitting slightly ahead of an 85W RTX 4060 and moderately ahead of an RTX 4050 is a coherent picture of a card operating well below its potential.

Against the HP Victus 15-fa1057ng (RTX 4050)

Across three shared games these two are level to within a frame — an RTX 4050 machine matching this one.

Here the sample size does the explaining. Three shared titles, drawn from the lighter end of what both machines have recorded, is the condition under which almost any two laptops of a similar era converge. At 1920x1080 in games with modest graphics demands, both machines are limited by how fast their processors can prepare frames rather than by how fast their graphics cards can draw them. The graphics card that has more capability in reserve has nowhere to spend it, and the result is a dead heat.

Set this against the five-game comparison with the Acer directly above and the difference in what they tell you is stark. The five-game tie is evidence about the hardware. The three-game tie is evidence about the games. Both are correct measurements; only one of them generalises. This is why we publish the number of shared titles alongside every margin, and why a buyer should always ask what a comparison was built on before letting it influence a decision.

Against the Asus ROG Strix Scar 17 (RTX 4090)

Across three shared games the PX13 averages 75fps less than the Scar 17. That is the widest margin in the comparison set, and it is worth pausing on, because it is a wide margin recorded on the same kind of narrow, processor-limited three-game sample that produced a tie in the paragraph above.

When a comparison built on lightweight titles still produces a 75-frame gap, the difference is not a measurement artefact. It means the two machines are far enough apart that even the games least suited to revealing the difference reveal it anyway. The Scar 17 is a 17.3in desktop-replacement machine with a flagship graphics card; the PX13 is a 13.3in ultraportable. They are not alternatives to one another in any practical sense, and the point of the comparison is calibration rather than shopping advice: it establishes the top of the scale so that the other margins can be read against something.

It is also a useful corrective to the temptation to read the PX13's better results — 217fps, 201fps — as evidence of a fast machine. Those figures come from titles that run fast on almost anything. Against genuinely fast hardware, on titles that discriminate, the ProArt is a long way back, and it was never designed not to be.

Reading the four together

The comparison set is unusually coherent. A moderate win over an RTX 4050 machine on five games, a dead heat with an 85W RTX 4060 machine on five games, a dead heat with another RTX 4050 machine on three light games, and a heavy defeat by a flagship on three light games. Every one of those results points the same way: the PX13 delivers upper-mid-range graphics performance, not the performance its graphics card name implies, and it delivers it in a chassis a third lighter than the machines it ties with.

That is a trade, and whether it is a good one depends entirely on what you value. Nobody else in this comparison set gets close to 1.389kg. Nobody in it has a 2880x1800 OLED panel either. If those are the things you are buying, the frame rate deficit is the price and it is a defensible one. If they are not, you are paying a substantial premium in performance-per-pound for portability you will not use.

Who it's for

This suits someone whose primary work is creative and whose gaming is secondary but real. That is a genuinely common profile and it is badly served by the market: most machines with this class of graphics card are large, loud and ugly, and most machines this small and this well-screened have no graphics capability at all. The PX13 sits in the gap. A 2880x1800 OLED, 32GB of memory and a 2TB drive make it a serious tool for photo and video work, and the graphics card means that when you close the editor, every game in this suite runs above the panel's refresh rate at 1080p.

It suits someone who genuinely carries their laptop. 1.389kg is the whole argument, and it is a strong one. If your machine travels daily, the difference between this and a 2.3kg gaming laptop will matter to you far more often than a frame rate difference will.

It suits someone with an external monitor. The 60Hz ceiling is the machine's main limitation as a games device, and it evaporates the moment you plug in a high-refresh display. A buyer who works at a desk with a monitor and travels with the laptop gets the best of both, and the substantial frame rate headroom in the lighter half of this chart finally has somewhere to go.

Who should buy something else

Anyone buying primarily to play games should look elsewhere. The reasoning is not that performance is bad; it is that you would be paying a portability premium for a machine whose panel cannot display most of the frames it produces. A larger laptop with a high-refresh screen, at a similar level of graphics capability, would give you a materially better gaming experience, and the comparison with the Acer Nitro V 16 shows that similar graphics capability is available in more conventional machines.

Anyone who wants to play modern titles at the panel's native resolution. The data does not support it. Cyberpunk 2077 at 86fps, Baldur's Gate 3 at 73fps and Horizon Forbidden West at 62fps are 1920x1080 figures, and 2880x1800 is more than double the pixels. Those titles will need a lower render resolution or upscaling to stay above 60fps, and you should plan for that rather than discover it.

Anyone who values high refresh rates. If you have played on a high-refresh display and would miss it, no other quality of this panel will compensate. That is a personal threshold rather than a technical one, but it is the single most common reason people are disappointed by an otherwise excellent screen.

Anyone who needs sustained graphics performance under long load. We hold no thermal or acoustic data for this machine, and a 13.3in chassis at 1.389kg has the least thermal headroom of anything in this comparison set. A machine that ties with an 85W RTX 4060 over five games is clearly working within firm constraints, and how it behaves in hour two of a session is not something the published averages can tell you.

The bottom line

The ProArt PX13 is a coherent machine that is easy to misjudge, because its graphics card name promises something the chassis was never going to deliver. Read it as an ultraportable creator laptop that can play everything in our suite above its own refresh rate, and it is impressive. Read it as an RTX 4070 gaming laptop and it will disappoint, because on five shared games it performs like an 85W RTX 4060 in a much bigger body. The panel is excellent in every dimension except the one gamers care most about, the weight is exceptional for the class, and the memory and storage provisioning is generous. We cannot score it because we have no confirmed UK price, but the performance data is complete, comparable and clear about what this machine is.

How it compares