%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
At 04:23 on April 23, Xiaomi made another move in the AI field, officially launching the new MiMo-V2.5 series large model and opening public testing at the same time. This release is not just a simple product update but a comprehensive breakthrough in the technical foundation. The series includes four models: MiMo-V2.5, V2.5-Pro, V2.5-TTS Series, and V2.5-ASR. Among them, the most attention from developers is that Xiaomi announced the global open-source of the core models MiMo-V2.5-Pro and MiMo-V2.5, demonstrating its ambition to promote an open AI ecosystem.
Flagship Performance: The "Terminator" of Millions of Context and Complex Tasks
As the top version of this series, MiMo-V2.5-Pro is positioned as a specialized tool for long and complex Agent (intelligent agent) tasks. It supports an ultra-long context of up to 1 million Tokens, showing amazing performance when handling complex software engineering and long-term logical tasks. To demonstrate its hard power, Xiaomi released a test case: during the development of the SysY compiler project for the Peking University course "Principles of Compilation," which originally required undergraduate students to spend several weeks, MiMo-V2.5-Pro completed it with 672 tool calls in just 4.3 hours and achieved a perfect score. This ability to maintain logical consistency over long-term tasks gives it the confidence to compete directly with top closed-source models like Claude Opus 4.6 and GPT-5.4.
Full-modal Evolution: Versatile and High "Value for Money"
Meanwhile, the standard version MiMo-V2.5 focuses on native full-modal capabilities, integrating image, audio, and video processing. In authoritative Agent evaluation rankings, MiMo-V2.5 not only significantly improves reasoning speed but also outperforms its previous Pro version. More exciting for developers is that through technical optimization, its API call cost has been reduced by about 50%. In cross-modal reasoning, video understanding, and professional chart analysis, MiMo-V2.5 is approaching or even surpassing some industry-leading models in certain metrics.
Revised Billing Model: Do More with Fewer Tokens
In addition to performance advantages, Xiaomi has also deeply optimized the underlying efficiency. Data shows that compared to competitors, the MiMo-V2.5 series uses fewer Tokens under the same testing benchmarks. To support the launch of the new model, Xiaomi fully upgraded the "Token Plan" subscription plan, completely eliminating billing differences between different context lengths and no longer implementing multiple rate billing. In addition, the newly introduced night-time discounted rates and auto-renewal mode further lower the entry barrier for enterprises and developers to enter the era of large models.