Very poor performance for x86 software. For a completely outrageous price.

In fact, any modern smartphone can easily be turned into a laptop with an external monitor and a keyboard + mouse, if the greedy smartphone manufacturers in all models did not disable the hdmi / dp ports built into the SoC via usb-c and usb-c itself was at least 5 -10Gbps. With parallel power supply to the smartphone from the PSU, and the most KEY - the operation of the battery power circuit as in laptops, i.e. when charging up to 100%, the battery of a smartphone that is being charged, like a laptop battery, will NOT be charged at the slightest drop in charge (which is just pure moronicity in all smartphones WITHOUT exception), but will turn on for recharging only upon reaching the THRESHOLD from which it is allowed to charge the battery. And the threshold MUST be adjusted by the user from 50 to 95%, as in laptops, where the maximum threshold is 95%.

It is this moronicity (intentional to come for a new model as soon as possible, because the batteries, when the smartphone is constantly on charge, are simply killed in less than a year) smartphone manufacturers do not allow you to surf the Internet from your smartphone in complete silence at home, simply by connecting it to usb- c a powered hub on a large monitor / TV + keyboard and mouse - even ordinary wired ones plugged into the hub.

This is all done on purpose, so as not to give universalism and the ability to work on a smartphone even without a battery, like a system unit, although it already has enough power. And this is done INTENTIONALLY so that there is no conflict with the market for cheap laptops.

And such options with such performance and such a monstrous price are absurd solutions. At a price of 3 times lower, maybe he would have been successful ...

Lots of interest from the Linux community for this model, would be great to see a review of that since distros run ARM natively.

Were the benchmarks run under x86 emulation? If so it is not very relevant to show them... It would have been better to compare with native benchmarks