MiMo-V2.5-Pro
Xiaomi’s open coding model, built to keep going through a thousand tool calls.
About MiMo-V2.5-Pro
MiMo-V2.5-Pro is Xiaomi’s flagship, a trillion-parameter model made for long, autonomous software work. Xiaomi’s showcase had it build a compiler in Rust over 4.3 hours and 672 tool calls, passing all 233 tests. It is also sparing with words: Xiaomi says it uses 40–60% fewer tokens than comparable frontier models to reach similar results.
It reads text only — the image and audio model is MiMo-V2.5 — and its weights are open under MIT. Xiaomi reports 57.2% on SWE-bench Pro.
Where it’s strong
Long-running autonomous coding sessions.
Agent work on a tight token budget.
Complex, multi-step software engineering.
Where to be careful
Text only.
Xiaomi’s leads are self-reported, and reviewers caution that such leads have not always held up in everyday use.
Xiaomi has not said how its reasoning can be controlled.
Price
For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost $2.83 — a long prompt and a page of answer each. Every rate is published and stays the same from one request to the next. See every model.
Specs
mimo-v2.5-proCall it
Use the OpenAI SDK you already have — point it at Roar AI and name the model. Read the docs.
from openai import OpenAI
client = OpenAI(base_url="https://api.roar-ai.com/v1", api_key="roar_live_…")
reply = client.chat.completions.create(
model="mimo-v2.5-pro",
messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)Compare MiMo-V2.5-Pro
Questions
How much does MiMo-V2.5-Pro cost?
Input is $0.5655 per 1M tokens; output is $1.131 per 1M tokens. For scale, 1,000 requests of about 3,000 tokens in and 1,000 out cost $2.83. The rate is published and does not change from one request to the next.
Can I use MiMo-V2.5-Pro alongside other models?
Yes. Every model on Roar AI is on the same API key and the same monthly invoice, so switching from MiMo-V2.5-Pro to another model is a change to one word in your code.
Is MiMo-V2.5-Pro open source?
Its weights are open under the MIT licence, so you could run it yourself. Through Roar AI you call it like any other model, with no servers to manage.