Music and sound · 2026-10-29

ElevenLabs upgrades its sound generator: SFX v3 builds whole sound scenes more reliably

ElevenLabs' new SFX v3 model creates sound effects from text descriptions, combining several sounds, music and vocals in one prompt, and is available now.

✨ What's new

ElevenLabs has introduced SFX v3, a new model that creates sound effects from text descriptions. According to the company, it follows prompts with several sounds more reliably than its predecessor SFX v2. In benchmark testing, it also produces more convincing music and vocals and cleaner audio without unwanted artifacts.

📌 Several sounds in one prompt

Previously, getting a scene with several sounds, such as a stadium where the crowd cheers and then boos, often required rewording the prompt many times. SFX v3 handles such descriptions better, so more of the sounds you describe end up in the result. You can start with the main action, add the surrounding sounds, listen to how they work together, and then refine the description.

🔹 Quality of individual sounds

The quality of each sound affects how natural a scene feels. A gong should resonate, a chime should ring cleanly, and a whistle should not sound harsh or crackly. ElevenLabs says SFX v3 is less likely to produce these unwanted artifacts, so the audio is closer to what you asked for.

🔸 Where to use it

You can create effects in ElevenCreative: in the Sound Effects tab, in Flows, in Studio, and in the Music Editor. There is also the ElevenLabs mobile app. Developers can access SFX v3 through ElevenAPI to add sound effect generation to their own products. For video creators, game developers and others working on projects, this makes it faster to find background sounds and effects.

Sources: 🔗 ElevenLabs Blog