MiMo V2.6 Flash
Neural Network
MiMo V2.6 Flash is Xiaomi's open Mixture-of-Experts model for text tasks and API integration.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 1 050 000 tokens
- Max answer length
- 131 072 tokens
- Reasoning effort
- low, medium, high
- Capabilities
- reasoning, function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works MiMo V2.6 Flash?
MiMo V2.6 Flash is Xiaomi's open Mixture-of-Experts model: out of 309B parameters, 15B are activated per token, and a hybrid attention mechanism ensures fast text processing. This allows the model to combine the knowledge base of a large system with efficient computation, making it suitable for streaming text tasks. In BotHub, it is available from Russia without a VPN or foreign card, and you pay with a Russian card in rubles for what you use — tokens that do not expire. Over 250 neural networks are available in one window, making it easy to compare answers, and developers have access to a unified OpenAI-compatible API: write the integration once and switch models directly in the request. Conversations are protected by AES-GCM encryption, and BotHub provides contracts, invoices, and electronic document management for companies, along with an admin panel and limits. Simple scenarios include: drafts of articles, emails, and descriptions, analysis and summarization of long documents, answering questions based on internal company materials, and supporting chatbots and services where fast text responses are crucial.Frequently asked questions about MiMo V2.6 Flash
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
This is a Xiaomi text model with a context window of up to 1,048,576 tokens. It is suitable for analyzing long documents, working with code, parsing images, and tasks requiring step-by-step reasoning and external tool calls. In BotHub, you can switch to another model in the same window.
Up to 131,072 tokens in a single response — that's dozens of pages of text. Enough for a large technical document, detailed code analysis, or a long translation. If the response hits the limit, ask it to continue: the model maintains conversation context and will finish the rest.
Yes, reasoning before answering is noted in our data. The model breaks the task into steps before providing a result, which is noticeable in math, logic, and code debugging. For simple questions, this mode is not needed; formulate your request briefly.
Yes, the model supports function calling and structured JSON output. Describe the available tools, and it will decide when to call them, returning the answer in the specified schema — convenient for agents and integrations. The BotHub API is compatible with OpenAI, so you won't have to rewrite your code.
Images and documents — yes: send a screenshot, diagram, photo, or file and ask it to analyze the content. Audio and video input are not noted in our data. You get text as output; image or sound generation is not noted for this model.
Our data does not specify prompt and response languages, so we cannot guarantee the quality of Russian. Test it with your task: in BotHub, the model is available without a VPN or foreign card, payment is in rubles, and other models are available nearby for comparison.