MiniMax H3 is at @anyscalecompute’s #RaySummit today in SF. @VictorSuOrtiz will break down our 33B open-weight audio-video model at the @vllm_project Conference, bringing text, image, video + audio into one context. @VoidAsuka will share how builders are fine-tuning, evaluating + serving H3 alongside @digitalocean, @nvidia, @inferact, and @NousResearch. From the model to the community building on it. See you there! 👋 🔗 Session details below:

Aug 26, 2026 · 9:16 PM UTC

9
4
96
19,043
Ray Summit: anyscale.com/ray-summit/2026 merge(happy_hour) → open_inference: luma.com/fufg0bg5

Ray Summit 2026 | Hosted by Anyscale

Join Ray Summit in San Francisco, Aug 24–26, for technical talks on foundation model training, multimodal AI, RL, and other AI in production systems.

anyscale.com
7
5,618
Sort replies: Relevant Recent Liked
Replying to @MiniMax_AI @anyscalecompute @VictorSuOrtiz @vllm_project @VoidAsuka
See you there! 🍻
1
384
Replying to @MiniMax_AI @anyscalecompute @VictorSuOrtiz @vllm_project @VoidAsuka
音视频统一进一个上下文，听起来挺猛，等实测了
181
Replying to @MiniMax_AI @anyscalecompute @VictorSuOrtiz @vllm_project @VoidAsuka
Great to see H3 bringing multimodal AI to the builder community.
6
Replying to @MiniMax_AI @anyscalecompute @VictorSuOrtiz @vllm_project @VoidAsuka
33B open-weight and it still shows up to the summit on time Ray Summit must be the one place where multimodal means showing up with slides
220
Replying to @MiniMax_AI
Let's Grow Send me Dm or Follow Back
99
Replying to @MiniMax_AI @anyscalecompute @VictorSuOrtiz @vllm_project @VoidAsuka
A 33B open-weight audio-video model handling text, image, video, and audio in one context is seriously exciting. As an AI creator, I’d love to get hands-on with MiniMax H3 Max and create some content around what it can do. Definitely keeping an eye on this!
144