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.

gen4-image-turbo is an image generation model from Runway. It works with a reference: you upload an image and add a text description, and you get a new frame as output. It is suitable for concepts, variations, visual refinement, promotional materials, and quick iterations on an existing image.

The output is an image, not text. Specific resolution and file format are not specified in our data, so we won't promise numbers: it's easier to evaluate the quality with your own task. The frame is built based on your description and the uploaded reference, so style and composition can be repeated from generation to generation.

Step-by-step reasoning is not noted in our data, and the task is different: it is an image generator, not a text model. You send a prompt with a reference and immediately get a frame. If you need text reasoning, reasoning models are available nearby in BotHub.

In our data, function calling and structured JSON output are not noted — the result comes as an image. However, you can access the model via the unified OpenAI-compatible BotHub API: keep orchestration and JSON schemas on the text model side, and call image generation as a separate step.

Images — yes, and according to our data, a reference is mandatory here: the model works in image-to-image mode, meaning it expects an image along with a text description. Audio and video input are not noted. The prompt context is 4095 tokens, which is enough for a detailed description of the frame.

Nothing is said about the prompt language in our data, so we won't make claims for the model — check it with your own task, it's just a couple of generations. But BotHub itself is fully in Russian: interface, support, payment with Russian cards in rubles, and without VPN.