Mistral Small 4
Neural Network
Mistral Small 4 is a Mistral AI model for reasoning and text tasks: access and Mistral Small 4 API in BotHub.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Mistral Small 4?
Mistral Small 4 is the next major release in the Mistral Small line by Mistral AI. Unlike previous versions, it combines the capabilities of several flagship vendor models into one system: the model blends strong reasoning with versatile text processing. In BotHub, it opens without a VPN or foreign card, and you pay in rubles only for the tokens you use, which do not expire. Alongside it, in one window, are over 250 models: you can easily compare answers to the same prompt and switch to another. A unified OpenAI-compatible API and batch mode provide a convenient way to connect the model to a multi-model system and process requests in batches without rewriting the integration. Chats are encrypted via AES-GCM, data is not saved, and companies have access to contracts, invoices, electronic document management, and an admin panel with employee limits. Use Mistral Small 4 for reasoning on complex questions, summarizing and analyzing long documents, extracting structured data from text, API-based conversational assistants, and mass processing of similar requests.Frequently asked questions about Mistral Small 4
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 Mistral text model: it answers questions, writes and edits texts, analyzes code, and works with input documents and images. A context of up to 262,144 tokens allows you to upload long contracts, reports, and repository fragments, while batch mode is convenient for processing large volumes of similar requests.
The output limit is up to 209,715 tokens, which is enough for a voluminous document, a detailed instruction, or a large code fragment in its entirety. In practice, you set the length yourself in the request or via API parameters, so standard responses remain compact rather than stretching to the maximum.
Yes, reasoning before answering is confirmed in our data: the model can break down a task step-by-step before providing a result. This helps with math, logic, code analysis, and long documents where it is important not to miss a detail and to double-check the output before the final formulation.
Yes, both. You describe the available tools, the model decides which one to call and returns arguments according to the specified schema, and the response can be obtained in strict JSON. This is convenient for agents, parsing into database fields, and integrations via BotHub's unified OpenAI-compatible API.
According to our data, in addition to text, images and documents are accepted as input: ask it to analyze a scan, diagram, table, or page from a file. Audio and video are not noted for this model, so it is better to choose a specialized model for them — in BotHub, it opens in the same window.
There is no separate evaluation for Russian in our data, so we will not make any guarantees: it is more reliable to test the model with your typical requests — a couple of prompts will show if the style and accuracy suit you. The BotHub interface is in Russian, access is without VPN or foreign cards, and payment is in rubles.