I probably mentioned it here before, about really "pushing" the technique of using the very common supported VT100 Query Cursor Position sequence to infer a lot of informations about terminals.
I published the results of 34 terminals with the capabilities detected, for Unicode WIDE+ZWJ+VS16 support that roughly translates to "does emoji", and LANG support for testing of hundreds of the worlds languages, I have also added column, 'Sixel' to report which of the 34 terminals tested support sixel graphics.
I also published a feature compatability chart of DEC private Modes for
these terminals. For developing rich CLI apps, you can take care to understand which terminals support features like synchronized output, bracketed paste, light or "dark mode", etc.
I hope this helps reconsider your "daily driver" when developing for modern terminals. Although I did test some legacy encodings, and even hope to look into developing with amiga terminal in 2026, but legacy "bbs clients" like syncterm do not support Unicode, so the "errors" logged is so great that it bogged down the builds and I regret that legacy terminals were excluded from results.
https://www.jeffquast.com/post/state-of-terminal-emulation-2025/
--- ENiGMA 1/2 v0.0.14-beta (linux; x64; 22.18.0)
* Origin: Xibalba -+- xibalba.vip:44510 (911:1801/0)