Key takeaways
- Xiaomi’s MiMo-V2.5 reportedly ranked first for token use over both one week and one month.
- China accounted for about two-thirds of tracked global calls to large language models.
- A token is a small chunk of text that an AI model reads or writes.
- High use shows demand, but it does not automatically prove one model is the smartest.
Xiaomi MiMo usage has reportedly reached the top of global AI token rankings. Xiaomi MiMo usage means the amount of text processed by Xiaomi’s MiMo-V2.5 model. A token is a small piece of text. Pandaily said China also handled about two-thirds of tracked large language model calls.
A large language model, or LLM, is AI that can read and produce text. Think of chatbots that answer questions or help write code. The figures point to a sharp shift in where people run AI tools.
What did Xiaomi MiMo usage lead?
Pandaily reported that MiMo-V2.5 ranked first worldwide for token use in both weekly and monthly measures. That matters because token use tracks real work done by a model. It can include a short chat, a long report, or thousands of lines of code.
The report separates two related measures. A call is one request sent to an AI model. Token use measures how much text moves through that request. One person might make 20 tiny calls, while another makes one huge call.
| Measure | Reported result | What it shows |
|---|---|---|
| Weekly token ranking | MiMo-V2.5 ranked No. 1 | Recent demand |
| Monthly token ranking | MiMo-V2.5 ranked No. 1 | Demand held over time |
| Tracked LLM calls from China | About two-thirds | China’s share of activity |
The clearest takeaway is this: MiMo-V2.5 was used heavily enough to lead the tracked token rankings, while China generated roughly two of every three recorded AI model requests.
Tracked global LLM callsReported share by location of model callsChina: about 67%Others: 33%MiMo-V2.5: No. 1 in weekly and monthly token use
Why is Xiaomi MiMo usage rising?
Price, speed and easy access can drive model use. So can a model’s skill at coding, maths, or Chinese-language tasks. The reported rankings do not prove which factor mattered most for MiMo-V2.5.
Chinese firms have also pushed hard to offer AI models at low cost. That can attract developers building apps for many users. Developers are people who make software and websites.
China has strong reasons to build its own AI stack. An AI stack is the full set of chips, data centres, models and apps behind an AI service. US limits on some advanced chip exports have added pressure, as our report on China chip export bans and the AI debate explains.
Does more token use mean a better AI model?
Not always. A popular model may be cheap, free for some users, or built into a busy app. It may also produce longer answers, which raises its token count.
That is why readers should not treat usage rankings like school exam marks. Tests of reasoning, coding and factual accuracy offer another view. Independent tests can help, but each test has limits too.
Still, a first-place ranking for both seven days and roughly 30 days is a useful signal. It suggests the demand was not just a one-day burst. Companies may now watch whether MiMo-V2.5 can keep that lead.
Why does China’s share of LLM calls matter?
About two-thirds is a huge share. If 100 tracked AI requests were placed in a day, around 67 came from China. That gives Chinese model makers a large home market to test products and gather feedback.
More use can help firms spot bugs quickly. It can also show which tools people want most. But use does not remove the need for safe design, clear rules and careful checks.
The race also reaches beyond chatbots. AI needs powerful chips and data centres, which are buildings packed with computers. Investors have already seen how chip demand can shake markets, including during the South Korea chip-stock sell-off.
What should users and businesses watch next?
First, watch whether Xiaomi MiMo usage stays high after the latest rankings. A monthly lead is stronger than a daily spike, but the market changes fast. New models can gain users within weeks.
Second, compare the model’s answers with other tools before trusting important work. Check facts, especially in school, health, money, or legal tasks. AI can sound sure even when it is wrong.
Finally, watch where the ranking data comes from. Public model platforms can show useful trends, but they do not measure every private company system. Readers can review OpenRouter’s public model rankings to understand how usage dashboards present model activity.
FAQs
What is Xiaomi MiMo-V2.5?
MiMo-V2.5 is Xiaomi’s large language model. It is AI designed to understand and create text, including answers and code.
How is Xiaomi MiMo usage measured?
Xiaomi MiMo usage is measured here through tokens processed. Tokens are small bits of text, so more tokens usually mean more AI work.
Why are LLM calls different from tokens?
A call is one request to an AI model. Tokens measure the size of that request and its answer, so one call can contain many tokens.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
