Realtime AI News
MiniMax updates Music3 music generation model on Hugging Face
MiniMax refreshed its Music3 model on Hugging Face, a text-to-audio pipeline for music generation and text-to-music tasks. The release runs on the sglang-omni inference library and ships with diffusers and safetensors support.
MiniMax updated its Music3 model on the official Hugging Face registry on August 13, with the listing showing a text-to-audio pipeline that turns text directly into audio content.
The official tags describe Music3 as a music-generation and text-to-music model running on the sglang-omni inference library, with diffusers, safetensors and pytorch labels indicating a standard open-source ecosystem format.
The model page currently shows 25 downloads and 1 like, a freshly registered version whose community traction is still in its early stage.
The Music3 naming marks the third generation of the music model line, and the official open-source listing on Hugging Face continues MiniMax's product-focused push into music generation.
Text-to-music is one of the hottest application areas of generative AI in 2026; compared with image and video generation, it adds clear value to creator toolchains and is widely seen as a key way for content platforms to cut material production costs.
What to watch next: whether Music3 will also go live on MiniMax's open platform with API access, and the specifics of generation quality, duration limits and commercial licensing terms.
For musicians and producers of video and podcast content, a text-to-music model update like this means lower creative barriers and lower production costs.
Why it matters
MiniMax's official open-source update of Music3 strengthens its position in music generation, with text-to-music capability poised to lower production costs for content creators.
Nearby Updates
All08/13, 23:30
Microsoft kills underperforming AI features and merges its separate Copilot apps
Microsoft is simplifying Copilot by combining its consumer and business apps, and dropping AI-generated podcasts, Group Chats, Deep Research and the Mico character, according to TechCrunch. The changes mark a retreat from experimental features toward a unified assistant experience.
08/14, 00:28
DeepSeek Releases DeepSeek-V4-Pro-0813, a New Model Registry Update on Hugging Face
DeepSeek has updated its Hugging Face registry with DeepSeek-V4-Pro-0813, a new text-generation model built on the transformers library and safetensors format. The MIT-licensed entry supports conversational tasks and endpoint deployment, and had already collected 174 likes at launch.
08/13, 23:08
Nvidia advances roughly $500B plan to keep aging GPUs valuable and unlock new AI financing
TechCrunch reported on August 13 that Nvidia is advancing a roughly $500 billion plan to keep aging GPUs from losing value and to persuade a new crop of financiers to keep lending for AI buildouts. The outlet calls the strategy risky but brilliant, noting it could inject massive capital into AI infrastructure while tying Nvidia more tightly to the credit cycle.
08/14, 00:51
California AI Bills Face Final Vote: Chatbot Safety, Copyright, US-First Commission
Multiple California AI bills are heading to a final vote today, covering chatbot safety, copyright, and the creation of a US-first AI commission, according to Tech Times. The outcome will shape the regulatory landscape for AI companies operating in the state.