MM Audio 2 Text-To-Audio (SFX)

Released in December 2024, MM Audio 2 is a specialized text-to-audio model for generating realistic sound effects from text. It prioritizes high-fidelity, specific sound events over continuous music, making it ideal for game development, animation, and film. Creators can produce custom SFX like footsteps or explosions without using pre-recorded libraries. The model interprets text prompts to synthesize a corresponding soundscape, offering a direct way to create needed audio assets.