Mistral Medium 3.5
Neural Network
Mistral Medium 3.5 is a Mistral AI model for code, agentic scenarios, and complex tasks, supporting text and image input.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Mistral Medium 3.5?
Mistral Medium 3.5 is a dense 128-billion parameter model from Mistral AI, trained to follow instructions precisely: it accepts text and images as input and outputs text, designed for agentic scenarios, programming, and complex multi-step tasks. The batch variant is intended for processing large sets of requests. In BotHub, the model is accessible without a VPN or foreign card; you can pay with Russian cards in rubles based on actual usage, and purchased tokens do not expire. Over 250 models are available in one window for easy switching, and a unified OpenAI-compatible API allows you to replace the model in a multi-model system without rewriting the integration, while chats are encrypted via AES-GCM. Developers will find code generation and refactoring, error debugging, and external tool calls within agent chains useful; analysts can read diagrams, tables, and screenshots; editors and marketers can create drafts and structure long materials; support teams can classify incoming tickets. BotHub provides contracts, invoices, and electronic document management for companies, and an admin panel with limits keeps expenses under control.Frequently asked questions about Mistral Medium 3.5
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 with a 262,144 token context. It analyzes long documents, writes and refactors code, answers questions about uploaded materials, accepts images as input, and calls external functions. Suitable for analytics, support, automation, and multi-model services via the unified BotHub API.
In a single response, the model outputs up to 209,715 tokens — enough for a voluminous report, a large code module, or a detailed document analysis. The total window is 262,144 tokens, so a long response should be planned together with the size of the request itself.
Yes, it has a reasoning mode: before the final answer, the model goes through intermediate steps. This significantly helps with math, logic, code debugging, and multi-step tasks where it is important not just to guess the result, but to arrive at it sequentially.
Yes. The model supports function calling and structured JSON output, making it easy to integrate into agents, chatbots, and pipelines: it calls your tools and returns data in a specified schema. In BotHub, this works via a unified OpenAI-compatible API, without rewriting the integration when switching models.
Images and documents — yes: you can send a scan, PDF, diagram, or screenshot and ask to extract data, compare, or explain. Audio and video processing is not noted in our data, so for voiceovers or video analysis, please select a specialized model from the catalog.
Mistral develops its models as multilingual, and in practice, they work confidently with Russian texts. There are no exact quality assessments for this specific version in our data, so it is most reliable to run a couple of your typical tasks and compare them with other models in one BotHub window.