by XiaomiReleased 21 September 2026
MiMo-V2.6-Flash is the efficient member of Xiaomi's MiMo-V2.6 line, and you can download it and run it yourself. It takes text, pictures, video and audio, writes back text, and keeps about 1,000 pages in mind at once. Xiaomi built it for long jobs it works through on its own, across coding, reading images, general questions and research.
You pay for what each request uses. No subscription, no minimum.
One message, the ordinary use
Every rate
| Meter | ↑ you send | ↓ comes back |
|---|---|---|
| Text tokens | 46,60 DA | 93,19 DA |
| Cached tokens read | 1,00 DA |
per 1M tokens
Eight companies run this model. OpenDunes sends each request to one of them, picked by price and speed. If it fails, the next one takes it. DA per 1M tokens: ↑ what you send · ↓ what comes back.
| Provider | ↑ Sent | ↓ Back | Cache read | Latency | Throughput | Uptime |
|---|---|---|---|---|---|---|
| Darkbloom | 33,28 DA | 93,19 DA | 16,64 DA | … | … | … |
| GMICloud20 % offUnited States | … | … | … | |||
| Cohere | 46,60 DA | 93,19 DA | 1,00 DA | … | … | … |
| Makora | 46,60 DA | 93,19 DA | 1,00 DA | … | … | … |
| DeepInfraUnited States | 46,60 DA | 93,19 DA | 0,93 DA | … | … | … |
| XiaomiAsia · Europe | 46,60 DA | 93,19 DA | 0,93 DA | … | … | … |
| VeniceUnited States | 58,25 DA | 116,49 DA | 1,25 DA | … | … | … |
| TenstorrentUnited States | 66,57 DA | 133,13 DA | 1,33 DA | … | … | … |
You pay the price of the provider that served your request. Prices read 15:06 · speed over the last 30 minutes · uptime over 3 days.
Real requests, real replies. Try any of them yourself — one click, pre-filled.
What's the total on this invoice, and when is it due?
invoice-scan.jpg · photoThe total is 42 300,00 DA, due 5 September 2026. Worth a double-check: the handwritten day could read as 8 instead of 3.
Try in Chat LabTurn this customer email into JSON with intent, urgency and order_id.
customer-email.txtTry your own request — same model, pre-selected.
Open Chat LabSwap the base URL and the key. Your OpenAI code runs unchanged.
import os
from openai import OpenAI
client = OpenAI(
base_url="https://opendunes.com/v1",
api_key=os.environ["OPENDUNES_API_KEY"],
)
resp = client.chat.completions.create(
model="xiaomi/mimo-v2.6-flash",
messages=[{"role": "user", "content": "Two lines to welcome a shop in Oran."}],
)
print(resp.choices[0].message.content)Runs this request once on your balance.
Standard OpenAI fields, grouped by what they control. Anything not listed is accepted and ignored.
Shaping the reply
How long, how varied, where to stop.
max_tokenstemperaturetop_ptop_kmin_pstopseedfrequency_penaltypresence_penaltyrepetition_penaltylogit_biaslogprobstop_logprobsReasoning
Let it think before answering — and see the thinking.
reasoninginclude_reasoningStructured output
Get JSON you can parse instead of prose.
response_formatstructured_outputsTools
Let the model call your functions.
toolstool_choiceparallel_tool_callsAbout 0,093 DA per message. ↑ 46,60 DA · ↓ 93,19 DA per 1M tokens.
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
MiMo-V2.6-Flash is the efficient member of Xiaomi's MiMo-V2.6 line, and you can download it and run it yourself. It takes text, pictures, video and audio, writes back text, and keeps about 1,000 pages in mind at once. Xiaomi built it for long jobs it works through on its own, across coding, reading images, general questions and research.
MiMo-V2.6-Pro-UltraSpeed costs 2,81 DA more for the same message.
Deposit dinars with your Edahabia or CIB card. Every model, MiMo-V2.6-Flash included, is paid from that balance. No foreign card, no conversion.
Xiaomi tags the model for English and Chinese. Arabic, French and Darja are not named.
It reads text, images, video and audio. Up to 1.1M tokens in one request, about 1 000 pages. It writes up to 98 000 words at once.
Yes, its weights are public. Open weights on Hugging Face
Xiaomi. Released 21 September 2026.
#6 for SVG drawings · MiMo-V2.6-Pro is Xiaomi's flagship, made for the most demanding work. It takes text, pictures, video and audio, writes back text, and keeps about 1,000 pages in mind at once. Xiaomi built it for long jobs it works through on its own, across coding, reading images, general questions and research.
An UltraSpeed edition of MiMo-V2.6-Pro built from the same trillion-parameter checkpoint, matching its quality while generating about ten times faster, for long agent tasks in coding, vision and research
MiMo-V2.5 takes in text, pictures, audio and video and writes back text, keeping about 1,000 pages in mind at once. Xiaomi says it handles multi-step jobs on its own as well as its Pro model while costing about half as much to run, and that it understands images and video better than the older MiMo-V2-Omni. Use it for work that mixes long documents with pictures or clips.
MiMo-V2.5-Pro reads text and writes text, keeping about 1,000 pages in mind at once and writing up to about 130 pages in one answer. It is Xiaomi's flagship for software engineering and long jobs it works through on its own, and Xiaomi says it can finish jobs that a team of specialists would spend days or weeks on, using the tools you connect more than a thousand times. Pick it for hard programming work.
From what its maker and two benchmark sites publish, and where it stands on OpenDunes. Read 7 Nov 2026.
Thinks first
Works through the problem before it answers.
Uses your tools
Can call functions your code provides.
Fills your JSON
Answers in the JSON shape you give it.
Searches the web No.
Can look things up online while it answers.
Cheaper on repeats
A prompt it saw a moment ago costs less the second time.
Open weights
The model's weights are public.
Run a coding agent across a long repository
Uses your tools
Answer questions about a recorded meeting
Audio
Describe what happens in a product video
Video