Models·Xiaomi

MiMo-V2.5: What It Does and What It Costs

xiaomi/mimo-v2.5

MiMo-V2.5 is Xiaomi’s open-weight AI model for text, audio, images, and video.

$0.14Input, per million tokens
$0.28Output, per million tokens
1.05MContext window
131KLongest answer

1.05M tokens of context — about 787,500 words, on a scale that ends at 2M.

What it can do

  • Reasoning
  • Vision
  • Tool use
  • Open weights
  • Takes text
  • Takes audio
  • Takes image
  • Takes video
  • Gives text

Released 22 Apr 2026 · checked 3 Oct 2026.

About it

MiMo-V2.5 is a Xiaomi model that accepts text, audio, images, and video, and responds with text. It supports reasoning and tools, has a 1,050,000-token context window, and its weights are open. Input costs $0.14 per million tokens and output costs $0.28 per million tokens.

Review

In a line

MiMo-V2.5 is an AI model from Xiaomi. It can take text, audio, images, and video as input, and returns text.

What it's good at

The model supports reasoning and tool use. Its inputs include images and video, so it can handle visual material as well as text and audio.

Its weights are open, meaning the model files are available for use.

Price in plain terms

Input costs $0.14 per million tokens; output costs $0.28 per million tokens. A token is a small piece of text the model reads or writes.

Reading a page of about 500 words costs under 0.01 cents. A chat turn with about 1,000 input tokens and 500 output tokens costs about 0.028 cents; 1,000 such turns cost about $0.28.

Context and limits

The model can take up to 1,050,000 tokens in a single context, roughly the amount it can consider at once. It can produce up to 131,072 tokens in one answer.

Its output is text.

Ways to use it

Its model ID is xiaomi/mimo-v2.5; its Hugging Face ID is XiaomiMiMo/MiMo-V2.5.

For more options, see open-weight models, Xiaomi models, and reasoning models.

Verdict

Consider MiMo-V2.5 if you want an open-weight model that can take audio, images, and video, and you need reasoning or tool use. Its input and output prices are low for the example workloads provided.

It may not suit someone who needs the model to return audio, images, or video, since its output is text only.

Verdict

MiMo-V2.5 is worth considering for people who need one model to read text, audio, images, and video, or who want reasoning and tool use with open weights. The example costs are small for the sample page and chat workloads. Its main constraint is output: it returns text only. Check that text answers and open weights fit your needs before choosing it.

  • Accepts text, audio, images, and video.
  • Supports reasoning and tool use.
  • Open weights are available.
  • 1,050,000-token context window.
  • Low input and output prices.
  • Returns text only.
  • Maximum answer length is 131,072 tokens.
  • Pick it ifPeople working with mixed text, audio, image, and video inputs; Readers seeking open weights with reasoning and tool use
  • Skip it ifYou need the model to generate audio, images, or video

Pricing

Input costs $0.14 per million tokens and output costs $0.28 per million tokens. The example costs are under 0.01 cents for a page, about 0.028 cents for one chat turn, and about $0.28 for 1,000 turns.

Questions

Who makes MiMo-V2.5?

Xiaomi makes MiMo-V2.5.

What can MiMo-V2.5 read?

It accepts text, audio, images, and video.

What does it produce?

Its output is text.

Does it support reasoning and tools?

Yes. The model supports reasoning and tool use.

How much does MiMo-V2.5 cost?

Input costs $0.14 per million tokens and output costs $0.28 per million tokens. A sample chat turn costs about 0.028 cents.

How much can it take in at once?

Its context window is 1,050,000 tokens, and it can write up to 131,072 tokens in one answer.

Are its weights open?

Yes. MiMo-V2.5 has open weights.

On these boards

More from Xiaomi