You are correct, my apologies!
Resolved it is :)

You're wrong, your code fails because you're using a global var that keeps its value between tests. Just reorder the sample tests to see it.

Respectfully, I think it is.
If you look closely, my code passed it too if I keep updating the sum in a loop.
If, in contrast, I collect all the values in a dictionary and then use Python's sum function on the values from the dictionary, I get a much larger number then the tests assert.
So, if the kata is trying to teach large number handling in Python, it is an issue, and hints would be appropriate.
If, on the other hand, it is just trying to teach algorithmic thinking, the tests should be changed to values where large number handling does not make a difference.
Best, Jovan

Not a kata issue. Before posting issues look at the top of the page: 11,143 people passed the Python version; if there were errors somebody would have seen them before you. Sorry and good luck!