NVIDIA Audex: unified audio-text LLM with zero text regression
NVIDIA released Audex (Nemotron-Labs-Audex-30B-A3B), a mixture-of-experts model that handles audio understanding, speech recognition, translation, text-to-speech, and audio generation in a single 30B backbone.
Built on Nemotron-Cascade-2, it preserves the text intelligence of its foundation model with negligible performance loss on text tasks.