MiniMax brings its 33B open-weight audio-video model H3 to Ray Summit
MiniMax_AI · x · 2026-08-27
MiniMax announced its presence at Anyscale's Ray Summit in SF: VictorSuOrtiz will break down the 33B open-weight audio-video model H3 at the vLLM Conference, covering text, image, video and audio in one context, while VoidAsuka will share how builders fine-tune, evaluate, and serve H3 alongside DigitalOcean, NVIDIA, inferact, and NousResearch.
Related event: MiniMax to Showcase 33B Open-Source Audio-Visual Model H3 at Ray Summit(2 posts)→
More from Companies & People
- Ex-OpenAI Leader: Human AI Researchers Have Maybe Two Years Left — 新智元 · 2026-08-27
- OpenAI employee: A few weeks at OpenAI feel like years elsewhere — soumitrashukla9 · 2026-08-27
- METR is hiring for Embedded Assessments technical staff to evaluate AI agents — CFGeek · 2026-08-27
- Commentary: The Next Nvidia is Nvidia, Not Cerebras — firstadopter · 2026-08-27
- Nvidia bears cross out peak year forecast again — firstadopter · 2026-08-27
- Zai_org founders are Tsinghua Yao Class professors with 20+ years AI experience — pstAsiatech · 2026-08-27