User demonstrates MiniMax H3 with image and voice references
Reddit user says they used 2 reference images and Kevin's voice clip, describing Scarlet's voice in one sentence, to produce an H3 MiniMax output.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Koharu translates manga pages locally using OCR, inpainting, and LLMs
- Orchestration tool runs AI coding agents in parallel, compares answers
- Real-time Gaussian splats generated on mobile devices
- Newtake AI shows off completely AI-generated rap music video
- ChatGPT Work used to install OpenClaw and Ollama, run local model