xAI: Grok Imagine Image 2.0

Grok Imagine Image 2.0 is an image generation and editing model from xAI, listed on api14 under the model id x-ai/grok-imagine-image-2.0. It sits in the image category and is built for two related jobs: creating images from text prompts and editing images from references. If you already route other xAI models through your stack, this one extends that line into visual output rather than text.

The model accepts text and image inputs and returns images. It has a context length of 65536 tokens and a maximum output of 58982 tokens. Generation is controlled through image parameters: resolution can be set to 1K or 2K, quality can be low or medium, and aspect ratio supports a wide set of values including 1:1, 3:4, 4:3, 9:16, 16:9, 2:3, 3:2, 9:19.5, 19.5:9, 9:20, 20:9, 1:2, 2:1 and auto. The n parameter accepts a value from 1 to 1, so each request produces a single image. Input references accept from 0 to 3 images, which is what makes reference based editing possible alongside plain text to image work.

Those capabilities point to a few practical patterns. A text prompt with no references covers straightforward generation, such as producing a single asset at a chosen aspect ratio and resolution. Supplying one to three input references shifts the model into editing and variation territory, where you guide the result from existing images rather than describing everything from scratch. The aspect ratio list is broad enough to cover square, portrait, landscape and tall or wide formats, so you can target specific placements without post processing crops. The low and medium quality settings let you choose between faster drafts and more finished output within the same model.

You call it through api14's OpenAI-compatible API using the model id x-ai/grok-imagine-image-2.0. The supported request parameters are logprobs, max_tokens, response_format, seed, temperature, top_logprobs and top_p. The seed parameter is useful when you want reproducible output across runs, and temperature and top_p give you control over how much variation you get between calls. Because the interface follows the OpenAI format, you can point existing client libraries at api14 and change the model id rather than rewriting your integration.

A few constraints are worth noting before you commit. Output is images only, so if you need text back from the same call, this is not the model for that. Each request returns exactly one image, so batch style workflows need multiple calls. Input references cap at three, which limits how much visual context you can supply for an edit. If your task is text generation or analysis rather than image creation and editing, a text model on api14 will fit better.