Google DeepMind Releases Lyria 3 Pro: AI Music Transitions from "30-Second Previews" to Full Songs

marsbitPublished on 2026-03-26Last updated on 2026-03-26

Abstract

Google DeepMind has launched Lyria 3 Pro, a significant upgrade from the previous version released just six weeks prior. The key improvement is the extension of AI-generated music from 30 seconds to 3 minutes, alongside a new "structure-aware" capability. This allows the model to understand and construct a song's internal composition—such as intros, verses, choruses, and bridges—based on user's text, image, or video prompts, marking a shift from a simple generator to a creative tool. While competitors like Suno and Udio have offered similar features since early 2025, Google's entry into this competitive space is notable due to its vast distribution reach via the Gemini ecosystem. The model supports multiple languages, includes vocals and instruments, and automatically adds SynthID watermarks for AI content identification. Access is tiered for paid Gemini users (AI Plus, Pro, Ultra), with daily generation limits, while free users remain on the 30-second Lyria 3. It is also available for developers via API and for enterprises through Vertex AI. However, copyright concerns regarding training data remain unresolved, similar to ongoing industry disputes.

Google DeepMind launched Lyria 3 Pro on March 25. Just six weeks after the release of the previous version, Lyria 3, this upgrade focuses on one core improvement: extending the generation duration from 30 seconds to 3 minutes, while enabling the model to truly understand the internal structure of a song.

This leap is not a minor iteration. Thirty seconds is sufficient for generating background sound effects but not enough to compose a song—there are no sections, no transitions, no climax. The newly added "structure-aware" capability in Lyria 3 Pro allows users to specify intros, verses, choruses, and bridges in their prompts, and the model arranges the transitions and dynamic changes accordingly. This marks a critical step for AI music tools evolving from "generators" to "creative tools."

Suno and Udio Have Been Doing This for a Year

Frankly, Suno and Udio already possessed this capability in early 2025, with both offering longer generation durations and more flexible structural control. Google catching up at this point signifies its true entry into competitive mode in the AI music race—backed by the distribution power of the Gemini ecosystem, Lyria 3 Pro will reach a much broader user base than any standalone music AI tool.

The simultaneous opening of Vertex AI is another signal: Google isn't just aiming for consumer tools but also plans to embed Lyria into enterprise workflows.

What It Can Do

Inputs support text, images, and videos, with the model automatically matching the music style to the emotional content. Generated content includes vocals, lyrics, and instruments, covering multiple languages. All outputs automatically embed SynthID watermarks to indicate AI origin—a consistent practice by DeepMind for content溯源.

Who Can Use It and How

Gemini App paid users can now access it. Usage is tiered by plan: AI Plus allows about 10 songs per day, Pro about 20, and Ultra about 50. Free users remain on the 30-second version of Lyria 3.

Supported languages include English, Japanese, Korean, Hindi, Spanish, Portuguese, German, French, and others, limited to users aged 18 and above. Access path: Gemini App → Create Music → Choose "Thinking" or "Pro" mode.

Developers can integrate via Google AI Studio and Gemini API; Vertex AI is in public preview for enterprise-level on-demand generation scenarios. Integration has begun with Google Vids and its music production tool ProducerAI. Enterprise Workspace support is expected within days.

Copyright Issues Remain Unresolved

Google states that the use of training data follows relevant agreements with artists but has not disclosed specific sources or authorization scope. This sits against the same backdrop of copyright lawsuits faced by Suno and Udio—legal disputes over AI music training data are still unresolved in the industry, and Google's statement is more of a position declaration than a complete answer.

Lyria 3 Pro is currently being rolled out gradually to users, with possible delays in some regions.

From Code to Cognition: A Ten-Thousand-Word Guide to the Evolution of the Robot Brain

"From Code to Cognition: The Evolution of Robot Brains" The journey of robotic intelligence has shifted dramatically from manually coded systems to AI-driven brains. For decades, robots relied on layered software stacks—perception, state estimation, planning, control—each handcrafted. While predictable, they lacked adaptability. The 2010s saw deep learning revolutionize perception (e.g., object detection) and control (via reinforcement learning), but learned skills remained narrow. The arrival of Large Language Models (LLMs) marked a turning point. LLMs acted as high-level planners, interpreting natural language instructions and generating sequences of actions for traditional robotic systems to execute. However, true integration came with Visual-Language-Action (VLA) models, which fused vision, language, and motion prediction into a single network. Pioneered by models like RT-2 and open-source projects like OpenVLA, VLAs enable robots to reason and act directly from visual input and commands. The most advanced humanoid robots now employ a "dual-brain" architecture: a slow-thinking, large VLA (System 2) for reasoning and planning, and a fast-reacting, small network (System 1) for high-frequency motion control, sometimes with an even lower-level System 0 for balance. This split balances cognition with the physics of real-time movement. Computation is split between onboard hardware (e.g., NVIDIA Jetson) for safety-critical control loops and cloud/edge servers for non-critical tasks like learning and interfaces. A crucial driver is the open-source ecosystem—models like GR00T and OpenVLA allow startups to build upon pre-trained brains and fine-tune them with their own data, accelerating development. Despite progress, current systems struggle with recovery from errors, sample inefficiency, and long-horizon tasks. This has spurred the rise of **World Models**—neural networks that predict the consequences of actions. By simulating possible futures before acting (like NVIDIA Cosmos or Meta V-JEPA), robots can plan, recover, and generalize better. This represents the next frontier: shifting intelligence from learned reactions to an internal model of physics and cause-and-effect. The field is rapidly evolving. While not yet at its "ChatGPT moment," the convergence of cheaper hardware, scalable simulation, and world models points toward robots that are increasingly capable, adaptive, and useful. The question is shifting from "what can robots do?" to "what *should* they do?"

marsbit4m ago

From Code to Cognition: A Ten-Thousand-Word Guide to the Evolution of the Robot Brain

marsbit4m ago

AI Bubble Is Bursting

The AI Bubble is Bursting: A Necessary Purge on the Path to Ubiquitous Intelligence Market volatility has reignited debates about an AI bubble, with figures like Ray Dalio pointing to high valuations. However, this parallels the dot-com bubble, which, despite its crash, laid the physical infrastructure for today's internet era. The current AI investment frenzy, with tech giants planning trillions in infrastructure spending far outstripping current AI application revenues, appears similarly imbalanced. This 'bubble' is seen as an inevitable phase for a disruptive technology, paying the "innovation tax." Critically, AI inference costs have plummeted over 99.7% since 2023, making intelligence nearly free at the margin. This hasn't reduced spending but has instead unlocked massive new demand, as seen in enterprise AI cloud expenditure tripling. This follows the Jevons Paradox: efficiency gains lead to greater total consumption. The market is now entering a cleansing phase, weeding out speculative ventures lacking real moats. The deeper shift is a move from capital expenditure (CapEx) on hardware to value creation in operational expenditure (OpEx) through AI applications that solve real industry problems. While infrastructure valuations are high, rapid earnings growth from widespread AI adoption across sectors—from manufacturing and finance to law and healthcare—may digest these valuations over time. Ultimately, this creative destruction will leave behind robust infrastructure and optimized models, cheaply powering an AI-augmented future for all industries, much as the internet became indispensable after its own bubble burst. The core productive potential remains undiminished.

链捕手14m ago

The AI Bubble is Bursting

The article argues that while a bubble in the AI sector is real and is currently showing signs of deflation, this is a typical, even necessary, phase for a transformative technology. It draws parallels to the dot-com bubble, where speculative excess was followed by the consolidation of internet infrastructure, enabling the digital age. Similarly, today's massive investments in AI infrastructure (data centers, power, cooling) far outpace current application-layer revenue, creating a valuation imbalance. However, the author contends this apparent bubble is not a sign of failure but of a transition. Key drivers include the plummeting cost of AI inference (over 99.7% in two years), which, per Jevons Paradox, has unleashed massive new demand as AI moves from simple tasks to complex agentic workflows across all industries. The market is now cleansing itself of shallow, speculative ventures, shifting value from capital expenditure (CapEx) on hardware to operational expenditure (OpEx) on transformative AI applications. The conclusion is that the underlying productive force of AI is undeniable. The current financial correction will eliminate weak players, leaving behind robust infrastructure and efficient models. This will ultimately fuel an "AI+" era where intelligence becomes as ubiquitous and essential as the internet, embedding itself into every sector of the economy.

marsbit15m ago

marsbit15m ago

The Next Bitcoin ETF Boom May Be Coming From Japan – Here’s Why

Despite recent outflows from US Spot Bitcoin ETFs, market attention is shifting to Japan as the potential next major market for such funds. This follows regulatory reforms in Japan that are moving cryptocurrency oversight to a financial instruments framework, potentially enabling Bitcoin ETF approval. Analysis suggests a Japanese spot Bitcoin ETF could attract significant inflows, ranging from ¥900 billion ($5.61 billion) in a conservative scenario to ¥3.1 trillion ($19.34 billion) in a bullish one. A base-case estimate of ¥1.4 trillion ($8.73 billion) would equate to demand for roughly 140,000 BTC. Beyond price impact, a Japanese ETF is seen as boosting Bitcoin's legitimacy and accessibility for investors and institutions. Bitcoin's price was around $61,038 at the time of writing, down 2.81% over 24 hours.

bitcoinist36m ago

The Next Bitcoin ETF Boom May Be Coming From Japan – Here’s Why

bitcoinist36m ago

Trump's 43 Minutes: The Strongman Narrative Unravels, the Media War Escalates

President Donald Trump's first public appearance in over a week, intended to project control amidst health concerns and political challenges, devolved into a 43-minute display of grievance and insecurity. He focused initially on a reflecting pool renovation, compared his rally size to Martin Luther King Jr.'s, and repeatedly attacked political opponents, judges, and U.S. cities. He signed executive orders, including one weakening job protections for senior federal employees, potentially replacing expertise with loyalty. His speech meandered to topics like communism before ending abruptly. The article highlights Trump's agitated, defensive state and the systemic pressures enabling his power. It details his personal attack on a CNN reporter, framing it as part of a broader assault on independent media to undermine public trust in unfavorable facts. This pressure is exemplified by the reported erosion of editorial independence at CBS. The author argues that as mainstream media faces political and commercial pressures, supporting independent journalists becomes crucial for preserving truthful information. The piece concludes by noting a symbolic House vote challenging Trump's Iran actions, suggesting growing Republican dissent, and frames financial support for independent media as a direct form of resistance.

marsbit4h ago

Trump's 43 Minutes: The Strongman Narrative Unravels, the Media War Escalates

marsbit4h ago

Trading

Spot

Futures

Hot Articles

Audiera: The AI Agent Network Powering the Web4 Entertainment Economy

Audiera is a dual-platform Web4 entertainment ecosystem combining a mobile rhythm experience and a lightweight Telegram mini-game, powered by AI interaction and an on-chain creator economy.

40.1k Total ViewsPublished 2026.03.11Updated 2026.03.11

Audiera: The AI Agent Network Powering the Web4 Entertainment Economy

The Cornerstone of the Autonomous AI Economy: How Talus is Reshaping On-Chain Intelligent Agents

Talus is a decentralized AI Agent framework built on the Sui, designed to solve the structural problems of current AI systems: centralization, opacity, and a lack of native economic identity.

42.9k Total ViewsPublished 2026.03.18Updated 2026.03.18

The Cornerstone of the Autonomous AI Economy: How Talus is Reshaping On-Chain Intelligent Agents

In-depth Analysis of AI and Crypto: The Era of Symbiosis between Algorithms and Ledgers

By 2026, the integration of artificial intelligence and cryptocurrency has advanced from proof-of-concept to a new stage of "system-level integration".

2.0k Total ViewsPublished 2026.03.26Updated 2026.03.26

In-depth Analysis of AI and Crypto: The Era of Symbiosis between Algorithms and Ledgers

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of AI (AI) are presented below.