ElevenLabs Music v2 & Stable Audio 3.0 Drop Major Upgrades

ElevenLabs and Stability AI unleashed significant new music generation capabilities on May 27, directly challenging Suno's dominance as the platform reportedly surpasses 100 million users. Both companies are spotlighting licensed training data in a clear nod to the legal heat still burning Suno and Udio from the 2024 RIAA suits.

🎚️ Breakthrough Features for Real Creators

ElevenLabs Music v2 delivers two killer features: seamless genre-switching inside a single track and true inpainting. Creators can now regenerate just a weak 8-bar hook or messy transition without rebuilding the entire song from scratch. This is the workflow upgrade professional AI musicians have been waiting for. No more discarding 45-second generations because the bridge fell flat. Early testers report it maintains coherence across drastic style changes, letting producers flip from trap to cinematic orchestral in one cohesive piece.

Stability AI countered with Stable Audio 3.0, shipping four models. Three come with open weights, enabling community fine-tuning and local experimentation. The small variants run on-device at 459 million parameters, perfect for low-power hardware like the new Coral boards. The medium model generates full 6:20 compositions in 1.31 seconds on an H200 GPU. Generation speed and length improvements close the gap with closed-source leaders while giving tinkerers raw models to build specialized genre tools.

💰 Business Moves and Ecosystem Impact

ElevenLabs, valued at $11 billion after its February Series D, slashed Music API pricing by up to 50%. The move lowers the barrier for studios and power users integrating audio generation into larger pipelines. By emphasizing exclusively licensed datasets, both firms position themselves as the "safe" enterprise options while Suno navigates expanded litigation.

For independent creators, this accelerates the shift from pure prompt-to-song toys toward professional production instruments. Inpainting fundamentally changes iteration speed. On-device models from Stability reduce reliance on cloud subscriptions and latency. Combine the two tools and workflows emerge that were impossible six months ago: rapid local prototyping on Stability followed by surgical edits and genre experimentation in ElevenLabs.

🔬 Technique Breakthroughs Worth Testing

Smart users are already experimenting with inpainting as a pseudo-stem editor, regenerating isolated elements while locking harmony and rhythm. Others are chaining genre switches to create evolving arrangements that feel human-composed. The open weights release from Stability invites custom LoRAs for niche sounds, echoing the explosion of fine-tuned image models in 2023-2024. Expect viral techniques around prompt anchoring for inpainting masks to spread quickly in Discord communities and YouTube tutorials over the next week.

These releases prove the AI music space is moving beyond basic generation into precision control and accessibility. Suno's user numbers show massive consumer adoption, but the pro creator stack is rapidly evolving with better tools, lower prices, and clearer legal footing.