Stability AI and ElevenLabs have launched competing AI music models designed to challenge Suno's market leadership, introducing features like longer track generation, section-by-section composition, and open-weight architectures. Stable Audio 3.0 generates tracks up to six minutes and 20 seconds with open weights on three of its four models, while ElevenLabs' Music v2 emphasizes compositional coherence across genre shifts and complex prompts.

What Are the Key Differences Between These New Music Models?

The two releases represent distinct technical approaches to AI music generation. Stability AI's Stable Audio 3.0 comes as a four-model family with varying capabilities and computational requirements. The Small models, designed for on-device use, contain 459 million parameters each and require no graphics processing unit (GPU). The Medium model reaches 1.4 billion parameters and can generate a six-minute-20-second track in approximately 1.31 seconds on an H200 GPU. The Large model, reserved for organizations with over $1 million in revenue, contains 2.7 billion parameters and is available only through Stability AI's API.

ElevenLabs' Music v2 takes a different path, focusing on what the company calls compositional coherence. According to the company, a single track can shift from opera to heavy metal and back while maintaining musical continuity, hold together through fast rap sections, and embed non-musical sound effects without falling apart. The model also introduces inpainting, which allows users to select a specific section of a song, regenerate just that part, and leave everything else untouched. Users can build songs section by section, starting with an intro, moving through verses and choruses, with the model maintaining continuity throughout instead of treating each clip as a standalone generation.

How Do These Models Address Copyright Concerns?

Both companies have made licensing deals central to their pitch, responding directly to copyright lawsuits filed against Suno and Udio by the Recording Industry Association of America in 2024. Stability AI has secured partnerships with Universal Music Group and Warner Music Group, while ElevenLabs has deals with Believe, Kobalt, and Merlin. These licensing arrangements represent a deliberate strategy to avoid the legal entanglements that have plagued other AI music companies.

For context, Udio settled with all three major record labels and is now a walled garden, meaning nothing users generate can leave the platform. Suno, by contrast, settled with Warner Music in November 2025, but Sony and Universal Music Group remain in federal court against the company. The licensing approach taken by Stability AI and ElevenLabs signals an attempt to build sustainable business models from the outset rather than face litigation later.

Where Do These Models Stand Against Suno?

Despite the technical advances in both Stable Audio 3.0 and Music v2, Suno remains the clear market leader. The company hit a $2.45 billion valuation in November 2025 and crossed $300 million in annual recurring revenue. Suno generates approximately 7 million songs per day and has been used by roughly 100 million people. These numbers represent a significant gap that neither Stability AI nor ElevenLabs has closed with their latest releases.

ElevenLabs, however, is building momentum in the broader AI audio space. The company reached a $11 billion valuation after a $500 million Series D funding round in February 2026 and hit $500 million in annual recurring revenue in April 2026. Music remains a small slice of that revenue, but ElevenMusic, which launched as a consumer app in April 2026, represents a direct competitive shot at Suno's user base.

Steps to Access and Use These New Music Models

• Stable Audio 3.0 Small and Medium: Available now on Hugging Face, an open-source model repository, allowing developers and creators to download and run the models locally on their own hardware without API fees.
• Stable Audio 3.0 Large: Accessible through Stability AI's API for organizations with over $1 million in annual revenue, providing the highest-quality output for commercial use cases.
• Music v2 for ElevenMusic: Free for ElevenMusic users, with commercial tiers available through ElevenCreative for brands and ElevenAPI for developers integrating the model into applications.
• Pricing Reductions: ElevenLabs cut Music v1 and v2 pricing by up to 50% for ElevenAPI and up to 40% for ElevenCreative self-serve, making the models more accessible to smaller creators and developers.

The technical architecture underlying Stable Audio 3.0 reflects Stability AI's commitment to developer accessibility. The company designed a semantic-acoustic autoencoder called SAME to hold melodic coherence over longer outputs. The model supports LoRA fine-tuning, which is a technique that allows artists to adapt the models to their own musical catalogs without retraining the entire model from scratch. Inpainting capabilities include single-segment, multi-segment, and causal continuation options, enabling users to extend a track past its original endpoint.

Stability AI has been technically credible in AI music for years without achieving major commercial breakthroughs. The open-weight strategy represents the company applying the same approach that made Stable Diffusion successful in image generation to the audio domain, seeding the developer community and observing what gets built. This contrasts with Suno's closed, proprietary approach and reflects a fundamental difference in business philosophy between the two companies.

The competitive landscape in AI music is shifting rapidly. ElevenLabs' price cuts and Music v2 launch signal aggressive expansion into a category where Suno has dominated. Stability AI's open-weight models and longer track generation address specific technical limitations of previous versions. Neither company has yet matched Suno's scale or user adoption, but the combination of improved features, cleaner licensing, and lower barriers to entry suggests the market may be entering a more competitive phase.