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.

The model works with text, analyzes uploaded documents and images, reasons before answering, and calls external functions. It is suitable for analyzing reports and screenshots, working with code, complex logical tasks, and scenarios where a structured response is needed for your service.

In a single response, the model outputs up to 32,000 tokens — enough for a long analysis, a voluminous document, or a large code fragment. The total context window is 65,536 tokens, which includes your request, attached files, and the response itself.

Yes, reasoning before answering is confirmed by our data: the model first processes the task internally and then provides the result. This significantly helps with mathematics, logic, code debugging, and multi-step questions where not only the answer but also the correct path to it is important.

Yes, the model supports both function calling and structured JSON output. You describe the schema or set of tools, and the response comes in a predictable format that can be passed on immediately. Through the unified OpenAI-compatible BotHub API, this can be connected without rewriting the integration.

Images and documents — yes: you can send a screenshot, diagram, PDF, or contract and ask to analyze the content. We have no confirmation in our data regarding audio and video input, so we will not promise them — it is easier to check directly in the interface with your own file.

The Gemini family was originally developed as multilingual, so the model usually handles requests in Russian confidently: it understands context and handles terms accurately. It is most accurate to check this with your own tasks; in BotHub, access is available from Russia without a VPN, and payment is possible with Russian cards.