I observed that increasing the idt period to 60uSec gave better results (more stable ) in this sim.
FYI, some improvements to the performance can be had by:
Tuning the output, similar eff%
590n inductor vs 3.3n cap is one option, 2nd harmonic => 23dBr down, 158W out
1975n inductor vs 800p cap is another => 34dBr down (2nd harmonic) @172W out
2500n inductor vs 625p = 2nd harmonic => 36dBr down, +/- 4pf tuning range, 165W out
3000n vs 515pf, 2nd harmonic 37dBr down
4000n vs 382p, 2nd harmonic 40dBr down now, +/- 2 pf cap tuning can vary amplitude & eff% a lot.
5000n vs 305pf, 2nd harmonic 43dB down, but capacitor tuning is +/- 1 pf now