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 an OpenAI Codex model tailored for working with code: writing new modules, refactoring, finding bugs, and explaining someone else's project. A 400,000-token context allows you to upload a large codebase along with documentation. It is also suitable for engineering tasks, document analysis, and agent scenarios.

The output limit is 128,000 tokens per response. This is enough for a voluminous module, a large refactoring, or a detailed analysis with comments without breaking the task into parts. The input context is 400,000 tokens, so the response takes the entire submitted project into account.

Yes, reasoning before answering is supported: the model builds a chain of thought before providing a result. Settings for reasoning effort and response detail are available — for simple edits, you can save tokens, and for complex architectural tasks, you can let the model think longer.

Yes, the model can call functions and return structured JSON according to a given schema — this is the foundation for agents that trigger your tools and APIs themselves. In BotHub, everything works via a unified OpenAI-compatible API, so you can switch to another model without rewriting the integration.

In addition to text, the model accepts images and documents as input: a screenshot with an error, an architecture diagram, or a PDF with specifications. We have no information about audio and video support, so count on text, images, and files — for media formats, BotHub has specialized models.

We do not have specific data on language support, so we won't make any promises — it's easier to test it with your own task. The BotHub interface, support, and payment with Russian cards are available without a VPN, and tokens do not expire, so a test request costs nothing extra.