AI & Tech

Google DeepMind unveils ‘Lyria 3 Pro’… Opening the era of longer and more structured AI music creation

AI composition model Lyria announces track length expansion and integration of all Google product lines with structure recognition function

AI Reporter Alpha··5 min read·
구글 딥마인드, '리리아 3 프로' 공개…더 길고 구조적인 AI 음악 생성 시대 열다
Summary
  • Google DeepMind unveiled 'Lyria 3 Pro', an AI music creation model equipped with structure recognition function.
  • This technology enables the creation of longer, more complete tracks and will be integrated into all Google product lines.
  • As quality competition in the AI ​​music market intensifies, discussions on platform integration and copyright are expected to accelerate.

Key announcement: Lyria 3 Pro released

Google DeepMind officially announced the next-generation AI music creation model 'Lyria 3 Pro'. The biggest feature of this version is the introduction of the Structural Awareness function. Unlike existing AI music generation models, which were limited to creating short clips or loops, Lyria 3 Pro understands the overall structure of a song, including intro, verse, chorus, bridge, and outro, and is now able to create longer tracks accordingly.

Google DeepMind also announced plans to expand the application of Lyria technology to Google's various products and platforms. This suggests the possibility that AI music creation functions will be integrated across the Google ecosystem, including YouTube, Google Workspace, and Android.

Why this is important: Breaking the ‘completeness’ barrier of AI music

The biggest limitations of AI music creation technology so far have been ‘length’ and ‘consistency’. Competitive services such as Suno and Udio can also create 2-3 minute songs, but they often show awkwardness in the narrative flow that runs through the entire song or the natural transition between phrases. The structure recognition feature is an attempt to address this problem head-on.

Another reason why this announcement is meaningful is Google’s platform strategy. If Lyria does not remain as a standalone service but is integrated into the Google ecosystem as a whole, it could have a significant impact on the creator economy. For example, it becomes possible for a YouTube Shorts creator to instantly create customized background music without a separate music license, or to automatically insert music that matches the mood into a Google Slides presentation.

Lyria model evolution process: comparison with previous versions

Lyria is Google DeepMind's music creation AI that was first released in collaboration with YouTube in November 2023. The initial version was introduced on a limited basis through YouTube's 'Dream Track' experimental feature, and has since developed into Lyria 2 and now Lyria 3 Pro.

ItemLyria 1 (2023)Lyria 2 (2024)Lyria 3 Pro (2026)
maximum track length30 seconds1-2 minutesMore than a few minutes (structure based)
Structure RecognitionNoneLIMITEDFull introduction
Platform integrationYouTube Dream Track LimitedPart of YouTubePlanned for all Google product lines
AccessibilityBy invitation onlybeta expansionScheduled for phased release

Although Google DeepMind did not disclose specific technical specifications (number of parameters, size of learning data, etc.), the introduction of the structure recognition function means that it goes beyond simple audio generation and reflects music theory and composition principles in the model architecture.

AI Music Market Trend: 2023 to 2026

2023 was the first year of popularization of AI music generation. Suno received explosive attention by unveiling its text-to-music service, followed by Udio. In 2024, the copyright debate will begin in earnest. Major labels such as Universal Music Group (UMG) have taken legal action against AI music services, and the issue of 'legality of learning data' has become a hot topic throughout the industry.

In 2025, the market entered the ‘quality competition’ stage. Beyond simply creating songs, the key was to produce results that could actually be used by professional musicians. The structure recognition function of Lyria 3 Pro comes from this context and shows that AI is developing in the direction of completing a ‘musical work’ rather than creating ‘one song’.

Market competitive landscape: comparison of major AI music services

serviceDeveloperCore StrengthsMain limitations
Lyria 3 ProGoogle DeepMindStructure recognition, Google ecosystem integrationBefore official launch, detailed specifications not disclosed
SunoSuno AIIntuitive UI, fast creation speedCopyright lawsuit in progress
UdioUdioHigh quality vocal synthesisCopyright lawsuit in progress
MusicLMGoogle ResearchText Prompt ComprehensionScheduled to be absorbed into Lyria
JukeboxOpenAISupports various genresDevelopment halted

[AI Analysis] Future prospects and implications

The emergence of Lyria 3 Pro is likely to affect the AI music generation market in three ways.

First, the acceleration of platform integration competition. If Google integrates Lyria into its product line, competing big tech companies such as Meta, Microsoft, and Apple are likely to make similar moves. Meta's AudioCraft and Microsoft's Copilot music functions could be strengthened.

Second, the copyright debate may enter a new phase. Since Google has a long history of cooperation with the music industry through YouTube, it is likely that it will enter into licensing agreements with labels and artists when commercializing Lyria. This can present a reference point for the ‘legal AI music creation’ model.

Third, there is a reorganization of the creator tool market. As content creators such as YouTubers, podcasters, and game developers are able to dramatically reduce music production costs and time, it is expected that existing royalty-free music library services will inevitably need to reexamine their business models.

However, as the exact release schedule, pricing policy, and technical specifications of Lyria 3 Pro have not yet been revealed, the actual market impact will likely have to be watched for future announcements.

Share

댓글 (3)

대전의돌고래1일 전

Google 관련 기사 잘 읽었습니다. 유익한 정보네요.

바람의별30분 전

간결하면서도 핵심을 잘 정리한 기사네요.

용감한토끼5시간 전

기사 잘 봤습니다. 다른 시각의 분석도 읽어보고 싶네요.

More in this series

More in AI & Tech

Latest News