llms ai-agents model-optimization developer-tools

MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

MiMo-V2.5-Pro-UltraSpeed: Pushing 1T-Parameter Model Generation Speed to 1000 TPS

mimo.xiaomi.com

June 8, 2026

8 min read

🔥🔥🔥🔥🔥

70/100

Summary

Xiaomi has released the MiMo-V2.5-Pro-UltraSpeed, capable of generating 1 trillion parameter models at a speed of 1000 transactions per second (TPS). This advancement enhances real-time AI reasoning and collaboration, making it more responsive and integrated into human thought processes.

Key Takeaways

Xiaomi released the MiMo-V2.5-Pro-UltraSpeed, achieving over 1000 tokens per second (tps) decode speed on a 1-trillion-parameter model for the first time.
The new model allows for parallel reasoning paths, enhancing the quality of AI-generated responses and enabling real-time decision-making in critical applications.
The MiMo-V2.5-Pro-UltraSpeed API is available through a limited-time application process, prioritizing enterprises and professional developers, with trial access from June 9 to June 23, 2026.
The model's speed significantly improves coding efficiency, allowing developers to generate code without the delays previously caused by inference latency.

Read original article

View story →

Coverage & discussions

Hacker News · mimo.xiaomi.com

MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second ↗

Community discussions

Hacker News discussion ↗

Historical count: 5 · 5 saved in sample · Historical sample

Analysis ready

Community Sentiment

Mixed

Positives

The introduction of MiMo-v2.5-Pro-UltraSpeed, achieving 1000 tokens per second, signifies a major leap in model efficiency, potentially transforming real-time AI applications.
The prospect of compute-in-memory technology could lead to models that are not only faster but also significantly smarter, enhancing the capabilities of AI systems.
Ultrafast AI could streamline workflows, allowing developers to focus more deeply on complex problems rather than getting bogged down in lengthy processes.
The ability to render UIs directly from context data and prompts could revolutionize software development, making it more accessible for small businesses.

Concerns

There are concerns that the rapid pace of AI development may lead to a proliferation of low-quality software, overshadowing genuine advancements.
The shift towards ultrafast AI might diminish the enjoyment of the craft for developers, as the focus shifts from deep problem-solving to quick fixes.
The increasing costs associated with AI usage, particularly from American providers, could create barriers for many companies, complicating the landscape for developers.
The inconsistency in model performance and pricing from AI providers raises concerns about reliability and predictability in AI applications.