NVIDIA Releases Audex (Nemotron-Labs-Audex-30B-A3B): A Unified Audio-Text LLM That Preserves the Text Intelligence of Its Backbone
NVIDIA introduces Audex 30B-A3B, a mixture-of-experts model combining audio understanding, speech recognition, translation, TTS, and audio generation while maintaining high text intelligence from its backbone.
- Audex 30B-A3B is a mixture-of-experts model combining audio understanding, speech recognition, translation, TTS, and audio generation in a single system.
- The model preserves the text intelligence of its Nemotron-Cascade-2 backbone with minimal regression.
- This release represents a step toward more unified multimodal AI architectures.
- Developers may benefit from reduced complexity in handling audio and text tasks.
NVIDIA has released Audex 30B-A3B, a groundbreaking mixture-of-experts (MoE) model developed by Nemotron Labs. This unified audio-text LLM integrates speech recognition, translation, text-to-speech (TTS), and audio generation into a single system. The model is built on the Nemotron-Cascade-2 backbone, ensuring that its text intelligence remains largely intact with minimal regression.
The release marks a significant step toward more cohesive multimodal AI systems, where audio and text processing are handled seamlessly within one architecture. By preserving the backbone's text capabilities, Audex 30B-A3B avoids the common pitfall of sacrificing performance in one modality for gains in another. This could streamline workflows for developers working with audio and text data, reducing the need for separate models for different tasks.
Source: NVIDIA Releases Audex (Nemotron-Labs-Audex-30B-A3B): A Unified Audio-Text LLM That Preserves the Text Intelligence of Its Backbone. Read the full piece at the source.
Simplifies integration of multiple audio-text capabilities into applications.
Potential for more efficient and cost-effective AI-driven audio-text solutions.
Advances the field of multimodal AI by unifying key audio-text tasks.
- Mixture-of-Experts (MoE)
- A model architecture where multiple specialized sub-models (experts) are combined, with a gating mechanism selecting the most relevant experts for each input.
GPT-5.6 Sol Review: Faster Coding, Half Fable 5 Cost, and a Benchmark Problem - Tech Times
Anthropic says Claude has carved out its own space to ponder - Axios
![[AINews] The Field Guide to Fable](https://images.weserv.nl/?url=substackcdn.com%2Fimage%2Fyoutube%2Fw_728%2Cc_limit%2F9fubhllmsBU&w=520&fit=cover&q=70&output=webp&dpr=2&we=1&il=1)
LLM[AINews] The Field Guide to Fable
tencent/Hy3
LLMCompose your agent prompts once, compile them to every harness
Model Panic: How Fear of Open-Source AI Is Ceding Ground to China - R Street Institute
The US is ceding ground to China in open-source AI due to fear and regulatory hurdles. This is allowing China to gain an upper hand in the development of AI technologies.
Meta Built An AI Detection Tool To ID Images And Video Created With Its New Models - Engadget
Meta has introduced an AI detection tool designed to identify images and videos generated by its latest AI models, aiming to combat misinformation and enhance transparency.
Kalshi traders see slim odds U.S. government will take a stake in OpenAI this year - CNBC
Kalshi prediction markets suggest less than a 5% chance the U.S. government will take a stake in OpenAI before 2025.
Forsyth County Sheriff's Office tests out new humanoid robot, artificial intelligence - wfmynews2.com
Forsyth County Sheriff's Office is piloting a humanoid robot equipped with AI for law enforcement tasks.
AI ToolsMeta rolls out Muse, a new AI image generator
Meta has introduced Muse, a new AI image generator designed for advertising and creator content. The model aims to streamline image creation for various use cases.
BusinessMeta Now Lets Anyone Use Your Instagram Photos in AI Images—Unless You Opt Out
Meta’s new Muse Image model can generate AI images using public Instagram photos unless users opt out. The change affects millions of creators.