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.

antigravity-preview-05-2026 is a Google model for working with text: it answers questions, writes and analyzes code, and analyzes attached documents and images. The context window is 65,536 tokens, so a large file or long correspondence fits into a single request. Batch request processing is supported.

In a single response, the model outputs up to 32,000 tokens — that's dozens of pages. It is enough for a voluminous document, a detailed analysis, or a large code file. If there is even more material, break the task into parts and ask it to continue the response in the next request.

A separate reasoning mode is not noted in our data for this model, but the lack of a label does not mean the capability is absent. In practice, the analysis is more accurate if you ask the model directly in the prompt to write out the solution steps and only then provide the conclusion.

Function calling and strict structured output are not noted in our data, so you shouldn't build your architecture around them in advance. Getting a response in JSON format is usually possible by describing the required structure directly in the prompt. Access is provided via a unified OpenAI-compatible API.

The model accepts images and documents as input and responds with text: you can give it a screenshot, diagram, table, or file for analysis. Audio and video input is not noted in our data. Image generation is also not claimed for this model.

We don't have data on languages for this model, so we won't make any promises — it's easier to check with your own real task, which takes a minute. The BotHub interface is in Russian, access works from Russia without a VPN or foreign card, and payment is made with Russian cards.