Google Magenta releases RealTime 2
Google Magenta has launched RealTime 2, an open-weights music generation model developed by Google DeepMind and optimized for low-latency streaming. Available on Hugging Face, the model generates audio faster than playback speed and supports multi-modal steering via text prompts, audio examples, MIDI, and hand gestures.
Google Magenta RealTime 2 transforms AI music generation from a batch processing task into a live, interactive instrument, bridging generative AI and physical performance.
* Low-latency streaming generation provides immediate auditory feedback, making it viable for live improvisation and adaptive game soundtracks.
* The inclusion of physical inputs (MIDI, gestures) makes the model more intuitive and accessible to musicians and non-musicians alike.
* The open-weights release encourages developer adoption and integration into creative tools.
DISCOVERED
2h ago
2026-06-04
PUBLISHED
2h ago
2026-06-04
RELEVANCE
AUTHOR
googleaidevs