NVIDIA Releases Audex-30B-A3B Audio Language Model
nvidia · hf · 2026-07-07
NVIDIA has released Nemotron-Labs-Audex-30B-A3B, a new model utilizing a MoE architecture (30B total parameters, 3B active). It focuses on audio-understanding and audio language modeling capabilities, supporting inference, SFT, and RL training. Positioned as a general-purpose reasoning model, it is open-sourced in transformers and safetensors formats.
Related event: NVIDIA Releases Open-Source Audio-Text LLM Audex-30B(9 posts)→
More from Models
- Reddit Users Say the Ming Image Model Is Seriously Underrated — alley103 · 2026-10-27
- GPT-6.1 Terra fills form in 53.5s vs 3:21 for Sol in computer use side-by-side test — sandersted · 2026-10-27
- DeepMind staff reply mentions "Gemini 4 models" in apparent early slip — vivnat · 2026-10-27
- OpenAI math repo adds new Lean-verified papers, including an ETH refutation claim — danintheory · 2026-10-27
- OpenAI models produce Lean-verified existence proofs for QCD and Yang–Mills — danintheory · 2026-10-27
- Codex ban update: users jailbreak AI agents to bypass Daybreak filter for reverse engineering — FlolightC · 2026-10-27