Andrew has a lot of dead air time in his livestreams. I like the "compact" reviews he does where that stuff is cut out. He does spend time playing games via crossover though so that is nice.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
M5 gaming
- Thread starter M4pro
- Start date
- Sort by reaction score
CP2077 benchmark
1200p Ultra - No Upscaling
MBP M5: 27FPS
MBP M4: 21FPS
From The Verge
1200p Ultra - No Upscaling
MBP M5: 27FPS
MBP M4: 21FPS
From The Verge
Last edited:
That is nearly a 30% uplift. Wonder if they will switch on RT/PT and see if the uplift carries.
“…games that have been specifically optimized for Apple silicon fare much better, performance-wise. For instance, I got Resident Evil 4 Remake to run at around 50 to 55 frames per second at 1080p resolution with MetalFX upscaling turned off in the game’s settings. With MetalFX enabled, the frame rate hovered between 100 and 120 FPS.”
From Tom’s Hardware
From Tom’s Hardware
“The SSD performance is outstanding. In Blackmagic Disk Speed Test, I consistently saw speeds over 6,000 MB/s for both read and write operations, sometimes topping 6,500 MB/s.”
“The chassis stayed comfortable, even during extended workloads. The aluminum enclosure does a good job of dissipating heat, and the underside of the laptop never got uncomfortably hot. This is a laptop you can actually use on your lap without discomfort, unlike some high-performance laptops that can get almost too hot to touch. Crucially, I didn't experience any thermal throttling. The M5 sustained its high performance throughout extended tasks without backing down. This consistency is what you need in a professional machine.“
From FStoppers
“The chassis stayed comfortable, even during extended workloads. The aluminum enclosure does a good job of dissipating heat, and the underside of the laptop never got uncomfortably hot. This is a laptop you can actually use on your lap without discomfort, unlike some high-performance laptops that can get almost too hot to touch. Crucially, I didn't experience any thermal throttling. The M5 sustained its high performance throughout extended tasks without backing down. This consistency is what you need in a professional machine.“
From FStoppers
“For real-world gaming performance, I turned to Lies of P, which also surprised me with some major leaps. With the M4 MacBook Pro, I could only get a steady 60 fps with the highest graphics settings in 1080p. With this M5 model, I was able to play at the highest resolution (3024 by 1890) between 70 and 75 fps. It was even smoother as I scaled down the resolution: The MacBook Pro hit 85 to 95 fps in 1,440p and up to 140 fps in 1080p.”
From Endgaget
From Endgaget
Notebookcheck shows a 44% uplift for CP2077 (no upscaling, or rt). 1080p Ultra.
www.notebookcheck.com
Test Apple MacBook Pro M5 (2025) - Der schnellste Single-Core-Prozessor der Welt
Notebookcheck testet das Apple MacBook Pro 14 M5 mit dem neuen M5-SoC, 32 GB RAM und Nanotexturglas-Mini-LED.
www.notebookcheck.com
Unfortunately… not in the current version of CP2077.
“But tossing in ray tracing makes it slower without adding anything. Cyberpunk's Low ray-tracing setting just beefs up local shadows a little. It's not worth it. So better, but not better enough.”
From CNET
Thanks. That result seems higher than I would have expected. I wonder what setting he tested at?
Dang they didn't update to the latest native CP2077 version (I was hoping but I knew a lot of their prior testing on older machines was using v2.2) nor use CB R24 for the energy test. Sigh. Oh well. Maybe they'll update it as time goes on.Notebookcheck shows a 44% uplift for CP2077 (no upscaling, or rt). 1080p Ultra.
View attachment 2570984Test Apple MacBook Pro M5 (2025) - Der schnellste Single-Core-Prozessor der Welt
Notebookcheck testet das Apple MacBook Pro 14 M5 mit dem neuen M5-SoC, 32 GB RAM und Nanotexturglas-Mini-LED.
However, 2.2 is good as a non-native test!
Could be the M4 Air was struggling with thermals?
His Blender result which he said was done with "no RT" was weird too - in the other direction. So not sure.
Also the screen on the MBP is mini-LED not OLED. Hmmmm ...
Last edited:
Definitely possible although the CP2077 benchmark is pretty short so I don’t know.
Lol. He called Blender, a rt benchmark, “no rt”. I assume he meant cpu and no hardware rt but still.
I think macOS only got 2.3, so any older version would be via Crossover. Unless the graph is just mislabeled.
Something I find particularly notable is that the m5 machine achieves around 50% of the RTX 5050 laptop from their database at x5.5 lower total system power consumption. That sounds pretty insane. What’s more, the power consumption is 10 watts lower than the M4, despite a massive performance boost.
Yeah it's crossover. They talked about Cyberpunk being native now in translated text, but the version they are using, 2.2, with all the historical data they can compare against previous models isn't native. Maybe they did do native for the M5, but then it isn't comparable, so ...
Interestingly NBC's Steel Nomad and SNL tests showed smaller gains than I was expecting. The M5 Max definitely won't be hitting the 5090 mobile in that test if that holds up.
Aye, nuts!
Aye. Also the uplift was small, like 25%? Pretty sure it's bigger than that from the other tests.
Thanks for picking this up, I hadn’t noticed. That’s disappointing. I get the reasons, but still.
Steel Nomad seems to be the outlier afaik. I wonder if SN is more simple than actual games and thus utilization is very high on pretty much all gpus? Given the lack of core count increase or frequency increase it’s not surprising. Just a guess.
Could also be that it runs at a "high" resolution with no upscaling. It is also more memory bandwidth sensitive (as is Solar Bay) than GPU clock speed sensitive.