Welcome to Beiming's AI Teahouse!

Right now, Suno has quietly rolled out its newest V5 model.

It's being celebrated as the most robust AI music - generation model currently on the market.

The timing of this launch is rather intriguing.

Just a short while ago, Universal, Sony, Warner, and a number of musicians set their sights on Suno. They accused the company of employing illegal "stream - ripping" technology to pilfer their copyrighted music from YouTube for AI training purposes.

So, this V5 release is rife with tension.

Let's first delve into the upgrades in this V5 version, which are claimed to be a quantum leap.

The most prominent highlight is, undoubtedly, the sound - quality upgrade, which has reached studio - level standards. The vocals sound more natural, with pronunciation, articulation, and emotion closely mirroring those of a real person. It even captures fine details like breathing.

Creative freedom has been elevated to a whole new level. Not only can you add vocals and accompaniment, but you can also break down a song into 12 separate tracks, such as drums, synthesizers, and vocals, just as you would in a professional workstation. This allows for easy replacement and reorganization.

Any inspiration that hits you can be instantly transformed into a song using its "Sample to Song" feature. Simply hum a few lines or record a guitar riff, and it can expand it into a complete song.

V5 also offers more robust control over music styles. Do you want the laid - back feel of jazz? The explosive energy of rock? The gentle touch of folk? Just one sentence will do the trick.

In the past, songs generated by music models had a very strong AI - generated feel.

This time around, V5 has raised the bar to the level of professional music production!

Netizens are also singing its praises. Let's take a listen to V5's capabilities together:

So, how can you get your hands on such an excellent model?

Well... by subscribing.

CTO Georg Kucsko put it straightforwardly in the promotional video:

V5 is currently only available to paid users because the new model requires a significant amount of computing power. Once we've ironed out the bugs, we'll consider making it more widely available.

So, for regular users, you're still stuck with V3.5 for the time being.

If you're eager to try V5, you'll need to subscribe to the Pro plan ($8 per month) or the Premier plan ($24 per month).

But there are some discounts available at the moment.

Despite unveiling the most powerful music model, Suno's journey isn't exactly smooth.

As mentioned earlier, a few days ago, the three major record labels filed a lawsuit, accusing Suno of illegally obtaining training data through "stream - ripping" — bypassing encryption to directly download audio from YouTube.

Moreover, independent singer Tony Justice spearheaded a class - action lawsuit, with the complaint ballooning from 24 pages to 68 pages.

It's worth mentioning that Anthropic was previously compelled to reach a $1.5 billion settlement with writers for training its model on pirated books.

This case serves as an excellent reference for music industry lawyers: while the use of training data may be debated under the principle of "fair use," if the data source is pirated, it's a clear - cut case of infringement.

For model companies, copyright has always been an unavoidable issue.

And for us as individuals, there are also some thought - provoking questions:

When AI - generated songs sound better than those created by humans, are you willing to forgo the "human touch," or will you still cling to those imperfect yet authentic works?

Feel free to share your thoughts in the comments section～