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 was created for working with interfaces: it analyzes a page screenshot and suggests the next action — click, text input, scroll, navigation. Plus standard tasks: text, document analysis, calling external tools. Suitable for agents, web scenario automation, and interface testing.

The maximum per response is 32,000 tokens, which is roughly several dozen pages of text. The context window is 200,000 tokens, so long correspondence, documentation, or a series of screenshots fit easily into a request, and the model keeps all of this in mind.

Yes, the reasoning mode is confirmed: before answering, the model builds a chain of steps. This is key for working with interfaces — it needs to understand what is on the screen and choose the next action. For complex multi-step tasks, the result is more accurate, though slightly slower.

Yes, both. The model calls functions and returns structured JSON according to your schema — it's convenient to connect it to your code and external services. In BotHub, this works via a unified OpenAI-compatible API, so you can swap in another model without rewriting the integration.

Images and documents — yes, this is confirmed: the model analyzes screenshots, photos, PDFs, and other files along with the request text. There is no note about audio and video in our data, so we won't promise them — it's easier to check directly in the BotHub interface with your own file.

Our data does not specify prompt and response languages, so we won't make any promises. The most reliable way is to test it with your own task: in BotHub, you send a request and immediately compare the result with another model, paying only for the tokens used.