Over time the 'accepted' values have changed. It is very hard to track down where various values used in your calculation actually come from (well, they are found in the appendices to Sze's Physics of Semiconductor Devices book).
As can be seen in, for example, Journal of Applied Physics, by 2003 the 'accepted' value for $n_i$ was $1.00 \times 10^{10}$ cm$^{-3}$, with that paper reanalyzing band gap narrowing to bring that down to $9.65 \times 10^{9}$ cm$^{-3}$. Such a value would require some or all of the 'accepted' values for $N_V$, $N_C$, and $E_g$ to change.
Essentially, this calls into question exactly how (and when) the various parameters listed by Sze (and other texts from the era) were determined, and why the discrepancy is not called out by Sze.
The 1991 JAP article cited in the review above does note that the "commonly cited value of $1.45 \times 10^{10}$ cm$^{-3}$" is "of uncertain origin."
In turn, the 1990 https://pubs.aip.org/aip/jap/article/67/6/2944/18297/Intrinsic-concentration-effective-densities-of article explicitly states about the intrinsic carrier concentration:
Despite the importance of these parameters, commonly used values ... are easily shown to form inconsistent sets.
and goes on with
The aim of the present work is to identify the source of the inconsistency mentioned above by critically evaluating the literature leading to the adoption of currently accepted values. As a result of this examination, a self-consistent set of these parameters was identified that removes the discrepancies previously noted.
