There's a new PlayStation 5 model on the market - a CFI-1100 series unit that replaces the launch model CFI-1000. I've already shared early impressions about the machine, finding that in all practical terms, there is no meaningful difference between PS5s old and new. However, questions remain about the decisions made in how Sony has delivered this new rendition of the PlayStation 5 and ultimately, if there really is any genuine difference between them, particularly in terms of longer-term implications. Crucially, if the cooling assembly has been the subject of a cost-reduction strategy, does the machine run hotter, and if it does, to what extent does that actually matter?
To assess the PlayStation 5 from all dimensions, I spent some time devising a series of performance tests for the console, comparing the new CFI-1100 model to a launch machine. This turned out to be somewhat more challenging than you might imagine because fundamentally, one of the biggest successes of the new wave of consoles is that game performance is generally excellent. Getting meaningful numbers is a case of isolating repeatable situations in a range of games where we can either unlock frame-rate, or bring about sustained drops beneath 60 frames per second. In theory, this testing should be entirely superfluous - because the whole point of a console is that all machines should run in exactly the same way. With that said, it occurred to me that PlayStation 5 does have a boost clock and while its implementation as described by Sony should ensure consistent results from one machine to the next, this has never been comprehensively tested. Meanwhile, some users erroneously believe the boost clock to be similar to a PC implementation, which does adjust frequencies according to temperatures - so why not put it to the test and put the whole matter to bed once and for all?
The second dimension to the testing comes down to the hardware make-up of the machine itself and after my initial report, I was approached by Steve Burke of Gamers Nexus to see about finding a machine for him to test. Gamers Nexus is renowned for the quality of its deep dive hardware analysis and nobody is more thorough in hardware testing, so I was happy to send on my unit - which Steve promptly stripped down to its barebones. The comprehensiveness of the work is such that even though other hardware reports have emerged since the original Austin Evans video that kicked off the controversy, Gamers Nexus' results are the most wide-ranging and exhaustive. I mean, Steve even created a 'Frankenconsole' PS5 to compare temperatures from the same board using both the original and revamped cooling assembly, eliminating potential differences caused by the silicon lottery (where no two chips that leave the production line are totally identical). In the process, we also learn more about the changes Sony made to the internals of the machine.
To the end-user though, the differences between PlayStation 5s old and new and minimal to say the least. The new machine is indeed lighter by around 300g, but the physical dimensions are the same and it's still a hefty unit. There is a small quality of life improvement, however, and yes, it's the now-infamous new screw for the stand. This allows for the stand to be attached easily enough by hand without the need for a screwdriver. If it encourages users to attach the stand, that's a good thing. In Gamers Nexus' tests, adding the stand gives a degree of additional cooling, although removing the side panels altogether is also advantageous.
Internally, the embedded video below from Gamers Nexus should tell you every you need to know about the performance of the new cooler, but perhaps the biggest takeaway is that Sony hasn't just swapped out the thermal assembly, also adjusting the mainboard itself and the baseplate. The bullet point takeaways from days of testing are relatively straightforward: there's an improvement to temperatures on the voltage regulators, memory temperatures are better in some respects and worse than others (but still only a few degrees difference overall) and while the main processor may well a few degrees hotter, there is no evidence that this presents anything worth worrying about, assuming you are keeping your PS5 in a well-ventilated area. As with any piece of modern gaming hardware, enclosing the console in something like a media cabinet is fundamentally a very bad idea, which is why I felt compelled to include some testing on this in my gameplay analysis on the new machine.
Elsewhere, the new PlayStation 5 has similar characteristics to the old one. I was intrigued to learn that fan temperatures are dictated by the power being drawn by the processor and not the actual temperature of the chip itself. I noted that the new PS5 seemed a touch louder than my launch units, but according to Steve at Gamers Nexus, this is entirely normal - and even between PS5s of the same generation, there can be a per-unit variance in fan speed of around 100rpm, explaining the difference. There's the sense then that the PS5 will run as hot it needs to in order to maintain system performance and curiously, even if it gets significantly hotter, the fans won't spin up. If the machine gets reaches unsafe temperatures, it'll automatically shut down instead. General use of the fan is curious - Sony is very conservative in its use and cooling could be improved with only a minimal change in acoustics. The overall takeaway is that ensuring a quiet experience appears to be Sony's top priority.
In terms of the gameplay tests, I tried to be as thorough as possible here, choosing a range of titles where we can reliably track - and repeat - performance drops from one PlayStation to the next. The only measurable variances I could find came down to the fact that one of my launch model PS5s was running beta firmware, whereas my CFI-1100 machine used retail firmware instead. Eliminating the beta firmware unit saw the new PS5 run pretty much identically to my other launch PS5. As of yesterday, all PlayStation 5 consoles updated to the latest system software and all variances disappeared. In approaching PS5 performance both in terms of analysing the boost clock and in testing the console running hot, I isolated areas in Godfall, Devil May Cry 5 Special Edition, Resident Evil Village, Control and Marvel's Spider-Man: Miles Morales where we could replicate game rendering that dropped beneath 60 frames per second.
Multiple runs were carried out to establish how stable these scenes were on the same machine (some areas of some titles would vary a touch from run to run) and then I compared to the CFI-1100. The end result is that the CFI-1100 ran exactly as we would expect - and for those concerned that the boost clock may drop back under high temperatures, I also put the new PS5 into a media cabinet with little airflow, ran it with the console consuming over 200W for two hours and then benchmarked it again, finding no difference whatsoever, despite the console itself being very warm to the touch when I retrieved it from the cabinet and being actually hot on the rear exhaust. The PS5 will keep on playing until it hits a specific max temperature and then it'll power itself down - but based on my tests, it won't get progressively slower in the way that, say, a PC CPU or GPU will.
So ultimately, I hope that this testing puts the matter of the new PlayStation 5 to bed once and for all, and the bottom line is that I can confidently stand by my initial test results. Yes, the cooler design in the new PlayStation 5 has been simplified in some respects and yes, the internal temperature results suggest that while cooler in some cases, the machine can run hotter in others. The core question of whether the new PS5 is better or worse than the launch model can be answered by saying that they're mostly much the same, certainly in terms of the end user experience. I've had a few people ask whether they should be actively looking to acquire an older console but ultimately, the new machine is fine, and that's good news bearing in mind how difficult it is to acquire any PS5 at all - let alone a launch model that has likely reached the end of its production run.
Will you support Eurogamer?
We want to make Eurogamer better, and that means better for our readers - not for algorithms. You can help! Become a supporter of Eurogamer and you can view the site completely ad-free, as well as gaining exclusive access to articles, podcasts and conversations that will bring you closer to the team, the stories, and the games we all love. Subscriptions start at £3.99 / $4.99 per month.