This may be a repeat as I haven't read the full thread. I don't know if it's as simple as "lying." The only tests sufficiently validated to be credible are the Wechsler and Stanford-Binet, neither of which go especially high, and neither of which have an appropriate sample size above ~2 standard deviations above the mean.
There are psychologists who claim to be capable of measuring IQ above this level, but they have no evidence to support their testing and their claims are implausible. Adult IQ is statistically defined and most tests use 15 as the standard deviation - such scales entail that a 160 IQ must be 4 standard deviations above the mean - most z tables don't even go that high, and the proportion of the distribution falling above 4 standard deviations is significantly fewer than 1 in 10,000. A 190 IQ, at 6 standard deviations above the mean, implies a frequency of roughly 1 in one billion. Since tests are standardized in nations with education systems, adequate nutrition, etc, that means significantly fewer than 7 people in the world actually have IQs in excess of 190.
Some tests use different values for the standard deviation. Stanford-Binet uses 16, and Cattel uses 22 (if I'm not mistaken). Regardless, no sample population has been sufficiently large to validate any of these extreme IQ scores. It's not even clear what it would mean to have such a high IQ score - even if g, the general intelligence factor, exists, measuring g directly may be impossible. Also, g may be limited by relatively diverse sets of skills and talents beyond a certain threshold and simple quantification of intelligence may be impossible. If there is no g, then it's meaningless to discuss "ultra-high" IQ scores in the first place.
So why do I say it's not as simple as "lying?" Because many people have taken so-called "IQ tests" that are nothing of the kind. For example, there is no such thing as a legitimate online IQ test, but many people are taken in and believe that some online test can accurately measure IQ. They aren't lying when they report absurd IQs in the 140, 160, or even higher (!) range, they have simply interpreted scams as legitimate tests and so believe their IQs are actually in these ranges.
A real IQ score is only relevant in context, so typically it's only useful to accept somebody's score if they mention the specific instrument used and the standard deviation that applies. A 140 on the Cattell B is much different from a 140 on the WAIS IV. There is no single "score" that applies, what matters is the number of standard deviations above the mean indicated by the test, along with the validity of the test.