Xiaomi MiMo-V2.5-Pro-UltraSpeed tops 1,000 tokens/s, running a 1T MoE model on a single 8-GPGPU node
Original title:🚀 1,000+ TOKENS/S ON A 1T MODEL! 🚀 We are thrilled to release Xiaomi MiMo-V2.5-Pro-UltraSpeed in ...
AI summary
Xiaomi MiMo, together with TileRT_AI, released MiMo-V2.5-Pro-UltraSpeed, achieving over 1,000 tokens/s output speed on a 1-trillion-parameter MoE model for the first time, using only a single standard 8-GPGPU node (not a Cerebras or Groq solution). A limited-time free chat experience is offered, with the UltraSpeed API priced at 3x and an output experience roughly 10x better. Applications run from June 8 to 23 (PDT), and enterprises can contact business-mimo@xiaomi.com by email.
Key Takeaways
- Xiaomi MiMo,together with TileRT_AI,released MiMo-V2.
- 5-Pro-UltraSpeed,achieving over 1,000 tokens/s output speed on a 1-trillion-parameter MoE model…
- A limited-time free chat experience is offered,with the UltraSpeed API priced at 3x and an output…
- Applications run from June 8 to 23 (PDT),and enterprises can contact business-mimo@xiaomi.
- com by email.
Quick Facts
- Published
- Updated
- Source
- X:小米 MiMo (@XiaomiMiMo)
- Category
- Model Releases
- AI score
- 82
小米用单节点8卡标准GPU在1T MoE模型上跑出1000+ tokens/s,没有走晶圆级或专用芯片的路子,直接把推理成本门槛拉低了一大截,做实时对话和Agent的可以申请免费聊天先上手感受一下。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Xiaomi MiMo-V2.5-Pro-UltraSpeed tops 1,000 tokens/s, running a 1T MoE model on a single 8-GPGPU node. June 8, 2026. https://new.nat6.net/en/news/cmq5bjjga06d1slt2lstyzq8i/