Why is Tesseract returning the wrong number here?

I'm using Tesseract to scan dates that have been cropped out of an image. A date will look e.g. like this:

For almost all dates this works perfectly fine, no problem at all. However, sometimes it trips up in a strange manner, usually involving 3s and 5s. In the case above, it sees the last 6 numbers as 0435000, which needless to say is 7 numbers, a 5 has been added in out of nowhere (as suspected, right after a 3). Another common mistake is interpreting "2300" as 3500. As far as I can remember it's always been only the last portion (the hours and minutes) that trips it up, never the first part of the date (I've been running it a lot of times daily for several months). Does anyone know what is causing this? And also perhaps what I can do to make it more consistent?

For what it's worth I'm not running Tesseract with any special commands, just tesseract date.jpg date.

The Image quality is very important and you should use rgb b&w:

import subprocess
import cv2
import pytesseract
# Image manipulation
# Commands https://imagemagick.org/script/convert.php
mag_img = r'D:\Programme\ImageMagic\magick.exe'
con_bw = r"D:\Programme\ImageMagic\convert.exe" 
in_file = r'Qkqxn.jpg'
out_file = r'Qkqxn_inv.jpg'
# Play with black and white and contrast for better results
process = subprocess.run([con_bw, in_file, "-resize", "70%","-threshold","35%", out_file])
# Text ptocessing
pytesseract.pytesseract.tesseract_cmd=r'C:\Program Files\Tesseract-OCR\tesseract.exe'
img = cv2.imread(out_file)
# Parameters see tesseract doc 
custom_config = r'--psm 7 --oem 3 -c tessedit_char_whitelist=_01234567890' 
tex = pytesseract.image_to_string(img, config=custom_config)
print(tex)
with open("numbers.txt", 'w') as f:
    f.writelines(tex)
cv2.imshow('image',img)
cv2.waitKey(0)
cv2.destroyAllWindows()

Output: