Xiaomi Extends MiMo-V2.5-Pro UltraSpeed Trial After 66,000+ Applications Flood In
Xiaomi announced today that it is extending the limited trial window for MiMo-V2.5-Pro-UltraSpeed, the high-performance inference mode that delivers up to 1,000 tokens per second. In a notification published on June 23, the company said the decision follows an unexpectedly massive wave of interest from the developer community.
Since its debut, MiMo-V2.5-Pro-UltraSpeed has attracted more than 66,000 applications from a broad cross-section of users — including numerous Fortune 500 companies, industry-leading enterprises, and individual developers — spanning fields such as law, finance, telecommunications, logistics, automotive manufacturing, media and culture, and higher education. The volume of applications, Xiaomi noted, has far exceeded initial expectations.
“We deeply believe that extreme inference speed will bring entirely new use cases and paradigms to the industry,” the company stated. “Seeing the community’s urgent and genuine demand, we want to give more users the opportunity to experience the paradigm shift enabled by 1,000 tokens/s.”
Xiaomi first unveiled the UltraSpeed mode on June 8, developed in collaboration with TileRT. The mode is positioned as a premium offering — priced at three times the standard MiMo-V2.5-Pro rate — promising roughly a tenfold improvement in output speed for API consumers. Approved users also receive limited free access to a Chat experience for trial purposes.
To maintain quality of service and fair usage under constrained resources, the trial comes with guardrails: each account is limited to 10 successful queue entries per day, individual sessions are capped at 30 minutes, and idle sessions are automatically released after five minutes of inactivity.
No firm end date has been set for the extended trial. Xiaomi says the shutdown timeline will be announced separately based on resource availability, with ample advance notice to allow users to prepare for migration and adaptation. New applications remain open for the duration, and previously approved users can continue using the service uninterrupted.