MiniMax-H3 omni-modal system ported to Apple Silicon via MLX

A new open-source Python package, MiniMax-H3-MLX, has been released to enable the MiniMax-H3 omni-modal generative system to run on Apple Silicon. This system can process text, images, audio, and video to generate up to 15-second video clips with audio. The package was successfully tested on an M5 Max MacBook Pro, requiring approximately 115 GB of model files and taking nearly 45 minutes for video generation. While the video output was impressive, the audio component produced unintelligible speech-like sounds due to a lack of specific audio prompting. AI

IMPACT Enables local generation of multi-modal content on Apple hardware, potentially increasing accessibility for creators.

RANK_REASON This is a port of an existing model to a new platform, not a novel model release from a frontier lab.

Read on Simon Willison →

• Apple Silicon
• huggingface_hub
• M5 Max MacBook Pro
• MiniMaxAI/MiniMax-H3
• PipeNetwork/minimax-h3-mlx
• pipenetwork/MiniMax-H3-MLX-8bit

• model release
• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →