Suno AI Just released v3.5 - Can it beat Udio AI? | AI Music Showdown
Based on MattVidPro's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.
Suno AI v3.5 can generate near–full-length songs in one run (roughly 1–2 minutes), reducing the need for multi-step stitching.
Briefing
Suno AI’s v3.5 update is a meaningful step toward matching Udio’s strengths because it can generate near–full-length songs in a single run—something Udio still struggles with, relying on short ~30-second chunks that require extensions. In side-by-side tests, Suno v3.5 routinely produced tracks around 1–2 minutes right away, and the overall song structure was noticeably more coherent than earlier Suno versions, making it easier to get a complete, listenable result without wrestling with multiple segments.
The most consequential difference showed up in structure and momentum. With a prompt for an “emo rap song about an alien searching for friends on Earth,” Suno v3.5 produced a track around 1 minute 38 seconds with improved arrangement and clearer progression into a beat. The reviewer still found Suno’s voice quality slightly behind Udio’s, but the gap narrowed because Suno’s v3.5 extension behavior and structure were more likely to “break into” a rhythm rather than staying stuck in an intro-like loop.
Udio’s output, by contrast, was judged stronger on vocal realism and overall sound depth. When the same alien prompt was run through Udio, the voices were perceived as better, and the sound felt more polished. However, Udio’s short initial generations made it harder to reliably land a full song with consistent structure—especially when extending. Even when Udio was extended, the reviewer repeatedly hit a problem: the track often didn’t fully commit to a clear beat-drop moment or a satisfying song “pickup,” leaving the composition feeling incomplete.
A workaround emerged for Udio: explicitly prompting “beat drop” during extensions. That injection improved results, but it still tended to skew toward the original segment’s direction rather than fully executing the requested structural change. When the reviewer compared a cherry-picked, high-performing Udio song against Suno’s recreation, Suno delivered a different but workable take—more consistent over time than earlier attempts, yet still slower and somewhat simpler in complexity.
Genre tests reinforced the tradeoffs. Suno v3.5 looked particularly strong for speed and length, producing a sad country folk track about an uneaten banana turning brown with a full, coherent feel in one go. Udio’s version sounded deeper, but the reviewer found it harder to continue and finish cleanly across multiple 30-second generations. Across both models, the conclusion was less about which one is “best” in every case and more about workflow: Suno v3.5 is the faster path to a complete draft, while Udio remains the more reliable bet for voice quality and richer sonic depth—at the cost of more iteration.
The update also hints at where Suno is heading next. While v3.5 is positioned as a bridge toward a forthcoming v4, Suno is also working on features beyond straight text-to-music, including turning uploaded audio clips into songs and generating music from sounds—an expansion that could broaden creative control and reduce the trial-and-error burden that currently comes with prompt tuning.
Cornell Notes
Suno AI’s v3.5 update is a practical upgrade because it can generate a near–full-length song in one generation (about 1–2 minutes), whereas Udio still starts with short ~30-second chunks that often need extensions. In head-to-head tests, Suno v3.5 showed improved song structure and a better chance of reaching the main rhythmic “pickup,” even if vocal fidelity still lags Udio slightly. Udio’s voices and overall sound depth were judged stronger, but continuing a song across multiple short generations could be inconsistent and time-consuming. The result is a workflow tradeoff: Suno v3.5 for speed and completeness, Udio for richer vocals and sound—especially when prompts are precise (like explicitly requesting a beat drop).
What is the biggest functional change in Suno AI v3.5 compared with earlier generations, and why does it matter for song creation?
How did Suno v3.5 and Udio differ on “song structure” in the alien emo-rap prompt test?
What strategy improved Udio’s results when the reviewer wanted a stronger beat-drop moment?
In the cherry-picked “Vegas/alien gambler” example, how did Suno’s recreation compare to Udio’s original?
How did the banana-turning-brown country folk test highlight each model’s strengths and weaknesses?
Review Questions
- In what way does Suno v3.5’s longer single-generation output change the difficulty of producing a complete song compared with Udio’s ~30-second chunk workflow?
- What evidence from the alien emo-rap test suggests Suno v3.5 improved structural momentum, even if vocal realism remained a weaker point?
- Why does explicitly prompting “beat drop” help Udio extensions, and what limitation still remains after using that tactic?
Key Points
- 1
Suno AI v3.5 can generate near–full-length songs in one run (roughly 1–2 minutes), reducing the need for multi-step stitching.
- 2
Suno v3.5 showed improved song structure and a better chance of reaching the main rhythmic “pickup,” compared with earlier Suno behavior.
- 3
Udio’s vocal quality and overall sound depth were judged stronger, but its ~30-second starting chunks make full-song continuity harder.
- 4
Adding explicit instructions like “beat drop” can improve Udio extensions, though results may still skew toward the original segment’s direction.
- 5
In genre tests (emo rap and sad country), Suno v3.5 favored speed and completeness, while Udio favored richer vocals and sonic depth.
- 6
The practical choice depends on workflow: Suno v3.5 for quick full drafts; Udio for higher-fidelity vocals and depth when time and iteration are available.