HUB probably tests 'open air' like most review sites do. That allows for better air circulation around the VRM area which will naturally be better for cooling.
He actually didn't! He used a high-end system for testing, though, so still essentially optimal conditions. You can find the testing setup here:
We test VRM thermal performance of 9 Intel Z690 motherboards priced between $220 and $300. If you're aiming for best value, these are budget boards to pair...
www.techspot.com
You mean that he used thermal vision sensor right? And how did he take this 60c from the left side of radiator or this one just above the socket? Or maybe it was an average temp from both ones?
According to his setup, he measured at the back of the PCB:
For recording temperatures we're using a digital thermometer with K-Type thermocouples. This allows us to report peak rear PCB temperature. but we're not recording Delta T over Ambient, instead we maintain a room temperature of 21 degrees and ensure a consistent ambient temperature with a thermocouple positioned next to the test system. For testing we've got three CPU configurations, excluding the F variants, it won't shock you that we've tested with the 12900K, 12700K and 12600K. The stress test consists in looping Cinebench R32 for an hour, at which point we're reporting the maximum PCB temperature, recorded using k-type thermocouples.
So they made sure to continuously check ambient room temperature, which was at 21°C, and used a very well-ventilated case with high-end CPU cooling. Thus, optimal testing conditions. You don't have optimal conditions, and there are many factors that influence VRM and other temperatures inside your case.
And my room temperatures are like... 26,5c so it maybe affects the VRM temps
Yes, it can definitely influence results. Also keep in mind that air flow inside the case is also important, since weaker airflow means that more hot air stays inside the influencing surrounding temperatures for the VRM as well. Also, please note that HUB called 75°C a "decent result" in the test I linked above, with a 12900k on ebtry level boards. If 75°C is decent under optimal conditions, then 68 under suboptimal ones is even better ^^