runtimeassertion

I'm so sorry to inform you that based on the photo of your inverter's sticker, this is an obvious clone as it has multiple missing signs! The error you've got is most probably due to hardware damage not related to the firmware update at all, just as Coulomb described earlier.
The model name says IV (4) not III (3) just so that you know the difference, because they're two separate models. The best thing to do now is to take it to the repair center, or much better to return the unit or change the model entirely.
You can refer to the following thread by @Coulomb for signs of clone inverters:
https://forums.aeva.asn.au/viewtopic.php?p=71752#p71752

I don't like to give recommendations, plus I'm not familiar with the GEW models. They seem to be an updated Axpert VM II, bit somehow they have RS-485 support for connecting to a BMS, without the usual removable or round display. I'm guessing that they have a separate card inside with the display processor, and that maybe drives the classic display, or maybe the DSP drives the display and the "display" processor only does BMS and comms.
So I'm not familiar with the firmware that such a model would have.
The 4024MT is a VM III Twin, so it presumably shares the same firmware as the 6 kW models. You get an extra 33% rated power from this model, which of course places higher demand on the battery. 4000W at 24 V is some 180 A, so you'd need a stout battery.

Sure, I found it in the forum here: It was said that it introduces smart fan control, and now I can really observe this.
 
Thanks for the explanations about these inverters. If I would have this information, probably I'd chose original version. I didn't have enough time, I needed the backup power urgently due to the critical situation in power grid and the stock was almost empty everywhere in the country..
As a software developer I wouldn't like if my software was stolen and sold under other names without any fees. I don't get why they use the same case color naming, is it legal at all to sell it worldwide, offering custom labeling... 
 
 

Hi @Coulomb,
Thanks a lot for the reply! I tried what you advised and it works now! 
Using your bypass cable directly flashing to the inverter, I started from 56.72 and it was succeeded. Then I tried again 60.89 and 68.04 - no luck (the process was failing at the middle with read error), so I reflashed on 56.72 and then tried 60.04. And it worked! Then I tried sequentially all 60.XX and ended with 68.04 successfully.
I couldn't imaging that if bootloader was the same the order of flashing firmwares matters.
Now I finally enjoy silent fans!
Hmmm, honestly I thought that all these inverters are OEM and available from different brands( Voltronic, Axioma, QPower, LogicPower etc) and were manufactured in the same facility somewhere in mainland China (OUYAD?), but now I see that originally it was from Taiwan, but then clones were produced  
But in any event, even if it is a clone, it looks robust and reliable inside.
 
Testing now in different modes...
 
Thanks again for help!

It looks like you are using 60.89 from the screen shot. I don't yet understand what the difference is between the 56.xx firmwares and the 60.xx firmwares. By the looks, it's something important 😲
I would first try to recover the DSP by reflashing with the latest 56.xx that you can find, probably 56.72.
Patiently retrying the reflash, making sure you have the right firmware file, possibly varying the timing from switch on of the inverter to pressing the update button on the reflash tool, as per my "if you think your inverter is bricked" post. 
Your inverter is a clone (note the cent symbols rather than the Greek letter phi for "phase" on the label). That reduces the chances of success. 
The bootloader is not changed by an ordinary reflash, successful or otherwise. The 56.xx and 60.xx firmwares use the same reflash command, so their target machines probably have the same bootloader.