What is up dudes and dudetss—Arlia here. In this video I dive into the brand new Suno v5.5 that dropped last night, and based on my initial testings it’s pretty damn good. I break down what Suno says are the three big pillars: Suno Voices (record/upload your own voice), Custom Models (train on your own original catalog), and “My Taste” (personalization that learns your genres and moods over time). The big takeaway from my side is the overall end result feels more authentic, especially the vocal performance—more expressive, with those little “eye candy” mannerisms that make it sound closer to a real singer. I also play a bunch of real generations: RPG/anime battle music, blended-language tracks (English + Japanese—still one of my favorite things because I’m a bit of a linguistic guy), some Arabic attempts, beatbox tests, and even an enka-style vibe. Then I test voice cloning using a non-singing sample (me reading a book), and it still manages to autotune and produce something surprisingly close—though I’ll be honest, the performance sounded kinda weird because the source sample wasn’t meant for singing. Finally, I compare an older v3.5 track to a v5.5 remake to highlight the jump in vocal expression, and I end with one clear limit from my test: I couldn’t get Mongolian throat singing to work yet.