Mistral Small 3.1 24B Instruct
Neural Network
Mistral Small 3.1 24B Instruct is a multimodal Mistral AI model: reasoning, coding, and image analysis via API.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Mistral Small 3.1 24B Instruct?
Mistral Small 3.1 24B Instruct is an updated version of Mistral Small 3 (2501) from Mistral AI: 24 billion parameters and multimodal input, allowing the model to process not just text but also images. It excels at reasoning, coding, and function calling. The main difference from the previous version is the added image understanding while maintaining the same compact size. Through BotHub, you can use it without a VPN or foreign card, pay with Russian cards in rubles for tokens as you use them, and they do not expire. A unified OpenAI-compatible API allows you to switch to any of over 250 models without rewriting your integration. Chats are encrypted with AES-GCM, and companies have access to contracts, invoices, electronic document management, and an admin panel with limits. It is useful for analyzing screenshots, diagrams, and document photos, writing and refactoring code in various languages, debugging, building agents with function calling, and handling large volumes of requests where a balance of quality and speed is important.Frequently asked questions about Mistral Small 3.1 24B Instruct
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: suitable for correspondence and editing, analysis and summarization, writing and parsing code, and working with documents. It also accepts images as input and describes their content. A 128,000-token context allows you to upload a large file in its entirety and conduct a long dialogue.
In a single response, the model generates up to 102,400 tokens — enough for a voluminous document, a long article, or a large block of code without splitting it into parts. If generation cuts off in the middle, simply ask it to continue, and the model will finish the rest.
A separate reasoning mode is not noted in our data. However, you can ask the model to break down the task step-by-step directly in the prompt: it will write out the solution process and then provide the result. For complex logic and calculations, this technique usually significantly improves the result.
Yes, the model supports function calling and structured JSON output, making it easy to integrate into agents, chatbots, and data parsing services. In BotHub, everything works via a unified OpenAI-compatible API, so you can switch models without rewriting your integration.
Images and documents — yes: send the file along with your question, and the model will analyze the diagram, screenshot, or contract and respond with text. Audio and video are not in the list of input formats, but BotHub has separate models for them in the same window.
There are no separate evaluations for the Russian language in our data, so it is more reliable to test the model on your own tasks. Write a couple of typical queries and compare the result with other models — in BotHub, they are available in one window, and switching takes seconds.