NVIDIA Releases Audex (Nemotron-Labs-Audex-30B-A3B): A Unified Audio-Text LLM That Preserves the Text Intelligence of Its Backbone

- NVIDIA dropped Audex, a unified audio-text LLM that actually keeps its text intelligence intact. Finally, a model that doesn’t suffer the usual "multimodal tax" where adding ears makes the brain forget how to read. It’s an open-source MoE beast handling speech, translation, and even non-speech sound generation without breaking a sweat on benchmarks. While moonboys are busy shilling garbage tokens, NVIDIA is quietly building the infrastructure for the next wave of AI interaction. No more fragmented stacks or regression headaches. Just pure, unadulterated capability. If you’re still relying on clunky, separate models for audio tasks, you’re operating like a relic. Adapt to the new standard or get left in the dust.