The theory of data transfer into Suno V6: what follows from the available description

The available description mentions Sony’s theory about the possible transfer of information between generations of Suno models and claims related to the use of copyrighted music. These details are not supported by the materials provided and have not been established by a court.

What was at issue in the described dispute involving Suno

The provided excerpt links the claims to the Suno V6 model and the question of whether copyrighted music could have entered it not only during initial training but also through the outputs of earlier models.

This is a separate question from whether the company used musical recordings for training. The text presents a more specific theory: information from copyrighted works may have passed into new models through generations produced by earlier versions. This concerns a position attributed to Sony, not a confirmed description of how Suno trained its models.

Details of the litigation, including procedural documents and the exact wording of the claims, are not disclosed in the information provided. The evidence submitted by the parties and the court’s assessment of individual incidents are also not described.

What is the theory of data transfer between models

According to the stated theory, copyrighted music may have influenced new models through the outputs of earlier generations. Under such a scheme, the original recordings would not necessarily be uploaded again into the training dataset for the new model: the assumption is that information was transmitted through the outputs of the earlier model, which could then have been used in developing the next one.

This is a description of a possible route of information transfer, not proof that Suno trained V6 in precisely this way. Confirming or disproving such a theory would require information about the datasets, the origin of user tracks, the preparation of training materials, and the links between development stages. The available facts contain no results of such a review.

It is important here to distinguish between a generated track’s similar sound, the use of a specific recording in training, and the transfer of information from one model’s output to another model. These circumstances concern different questions and do not, by themselves, confirm one another. A mere description of the alleged chain is insufficient to establish copyright infringement.

What the provided description says about training V6

The text states that Suno trained V6 from scratch on a new dataset that included user tracks. The materials provided contain no primary company publication confirming this statement. In addition, the wording does not disclose the origin of each track or answer whether the dataset was connected to the outputs of earlier models.

What exactly does this description call user tracks: songs uploaded by users, music created in the service, or both types of material? Without clarification of the dataset’s composition, the phrase allows several interpretations. It also does not show whether generations from previous versions were included in the prepared data.

The phrase “trained from scratch” does not, by itself, resolve this question. It may describe the launch of a new training stage; however, the reported information contains no technical definition showing exactly what data were used or whether outputs from earlier models were excluded. Therefore, the statement attributed to the company and the plaintiffs’ accusation should be considered separately.

How does the described dispute differ from the reference to the 2024 case

The provided text also refers to a 2024 dispute involving the use of music to train models. The materials presented do not confirm the existence of this proceeding, its participants, or the substance of the claims. Therefore, this reference remains an unconfirmed report.

The described theory emphasizes the possible transfer of data between models. It does not allow one to conclude that the referenced dispute has ended or that the new lawsuit replaced the old one. The available information is insufficient for such conclusions.

For musicians and rights holders, the distinction concerns the source of the training materials and the chain of their subsequent use. Suno users should also avoid conflating the service’s creation of a track with proven use of a specific copyrighted recording in training: the former, by itself, does not establish the latter.

What is known and what remains the parties’ position

The information provided contains a theory that information from copyrighted works may have entered a new model through the outputs of earlier models. Information attributed to Suno about training V6 is also included in the text, but the primary source of this statement has not been provided. The description contains no data that would establish whether such a transfer occurred or what materials were used at each stage.

• Reported: in the provided excerpt, the claims against Suno are linked to the V6 model and the use of copyrighted music.
• The stated theory: information from protected works may have entered the new model through the outputs of earlier models.
• The statement attributed to Suno: V6 was trained from scratch on a new dataset that included user tracks; the primary source of this wording has not been provided.
• Unclear: whether user tracks meant music uploaded by users, Suno generations, or both types of material.
• Not established: whether the described data transfer occurred and whether Suno infringed copyright.

The stated theory and the plaintiffs’ allegations are not equivalent to a court decision. The available description contains no judicial finding that Suno used specific copyrighted recordings to train V6 or transferred information between models. It is more accurate to speak of the position presented in the text and the lack of data, rather than of proven infringement.

Frequently asked questions about the described dispute involving Suno and the V6 model

Does the stated theory link Suno to data transfer between models?

Yes. The text says that information from copyrighted works may have passed into new models through the outputs of earlier models. This is a theory attributed to Sony and the plaintiffs, not an established fact.

What does the text say about the V6 dataset?

It states that Suno allegedly trained V6 from scratch on a new dataset containing user tracks. The primary source of the statement has not been provided, and the available information does not clarify whether these were user uploads or songs created in Suno.

How does the described theory differ from the reference to the 2024 case?

In the text, the reference to the 2024 case concerns the use of music for training. Separately, it describes the possible transfer of information from earlier models into V6. There are no materials confirming the substance or status of the 2024 proceeding; this reference remains unconfirmed.

Has copyright infringement been proven?

No. The available information describes claims and statements attributed to the parties, but contains no judicial finding of infringement or confirmation of the alleged data-transfer chain.