--- title: Academia | Scenario Docs --- > This page is auto-generated from model configurations. Last updated: 2026-04-16. This reference lists all available **Academia** image generation models and their parameters. Use these parameter names when calling the [Generation API](/api/postgeneratecustom/index.md). - [Joyai Image Edit](#joyai-image-edit) - [Physic Edit](#physic-edit) - [Z-Image](#z-image) - [Z-Image Turbo](#z-image-turbo) --- ## Joyai Image Edit All-in-one image editing model with strong visual understanding for precise natural language edits. **Model ID:** `model_joyai-image-edit` **Capabilities:** `img2img` **LLM Markdown:** | Parameter | Type | Required | Default | Min | Max | Allowed Values | Description | | ------------------- | ------ | -------- | ------- | --- | ---------- | -------------- | ------------------------------------------------------------------ | | `prompt` | string | Yes | - | - | - | - | The edit instruction describing what changes to make to the image. | | `image` | file | Yes | - | - | - | - | Input image to edit. | | `negativePrompt` | string | No | - | - | - | - | Text describing what should not appear in the output image. | | `numInferenceSteps` | number | No | `30` | 1 | 50 | - | Number of denoising steps. | | `cfgScale` | number | No | `5` | 1 | 15 | - | Controls how closely the output follows the edit instruction. | | `seed` | number | No | - | 0 | 2147483647 | - | Random seed for reproducibility. | ## Physic Edit Physics-aware image editing with realistic refraction, material changes, and deformations. **Model ID:** `model_physic-edit` **Capabilities:** `img2img` **LLM Markdown:** | Parameter | Type | Required | Default | Min | Max | Allowed Values | Description | | ------------------- | ------ | -------- | ------- | --- | ---------- | -------------- | --------------------------------------------------------------------- | | `image` | file | Yes | - | - | - | - | Input image to edit with physics-aware transformations. | | `prompt` | string | Yes | - | - | - | - | Editing instructions describing the physical transformation to apply. | | `negativePrompt` | string | No | - | - | - | - | Description of unwanted elements in the output. | | `numInferenceSteps` | number | No | `40` | 1 | 50 | - | Number of denoising steps. | | `cfgScale` | number | No | `4` | 1 | 20 | - | Classifier-free guidance scale. | | `seed` | number | No | - | 0 | 2147483647 | - | Random seed for reproducibility. | ## Z-Image Z-Image - Highest quality text-to-image and image-to-image generation **Model ID:** `model_z-image` **Capabilities:** `txt2img`, `img2img`, `controlnet` **LLM Markdown:** | Parameter | Type | Required | Default | Min | Max | Allowed Values | Description | | ----------------------------- | ------------- | -------- | ------- | --- | ---------- | -------------- | --------------------------------------------------------------------------------------------------------------------------------------- | | `modelId` | model | No | - | - | - | - | A single LoRA model ID. If provided, it will be added to the loras array. | | `loras` | model\_array | No | “ | - | - | - | List of one or more LoRA model IDs. If modelId is a LoRA and this is not provided, it will be automatically populated with the modelId. | | `lorasScale` | number\_array | No | - | 0 | 2 | - | Scales for the LoRA weights | | `prompt` | string | Yes | - | - | - | - | Text prompt for image generation | | `negativePrompt` | string | No | - | - | - | - | Text prompt for what to avoid in the image | | `image` | file | No | - | - | - | - | Source image for img2img transformation | | `strength` | number | No | `0.6` | 0 | 1 | - | How much to transform the reference image. 0 means no change, 1 means complete transformation | | `controlImage` | file\_array | No | - | - | - | - | Control images to use as input (edge, depth, or pose) | | `controlnetConditioningScale` | number | No | `0.75` | 0 | 1 | - | The scale of the controlnet conditioning | | `numOutputs` | number | No | `4` | 1 | 4 | - | Number of images to generate | | `numInferenceSteps` | number | No | `28` | 1 | 100 | - | Number of inference steps. | | `width` | number | No | - | 128 | 2048 | - | The width of the generated images, must be a multiple of 16 | | `height` | number | No | - | 128 | 2048 | - | The height of the generated images, must be a multiple of 16 | | `guidance` | number | No | `5` | 1 | 10 | - | Controls how closely the generated image follows the prompt | | `seed` | number | No | - | 0 | 2147483647 | - | Used to reproduce previous results | ## Z-Image Turbo Z-Image Turbo - Fast text-to-image and image-to-image generation model **Model ID:** `model_z-image-turbo` **Capabilities:** `txt2img`, `img2img`, `controlnet` **LLM Markdown:** | Parameter | Type | Required | Default | Min | Max | Allowed Values | Description | | ----------------------------- | ------------- | -------- | ------- | --- | ---------- | -------------- | --------------------------------------------------------------------------------------------------------------------------------------- | | `modelId` | model | No | - | - | - | - | A single LoRA model ID. If provided, it will be added to the loras array. | | `loras` | model\_array | No | “ | - | - | - | List of one or more LoRA model IDs. If modelId is a LoRA and this is not provided, it will be automatically populated with the modelId. | | `lorasScale` | number\_array | No | - | 0 | 2 | - | Scales for the LoRA weights | | `prompt` | string | Yes | - | - | - | - | Text prompt for image generation | | `negativePrompt` | string | No | - | - | - | - | Text prompt for what to avoid in the image | | `image` | file | No | - | - | - | - | Source image for img2img transformation. Cannot be used with ControlNet. | | `strength` | number | No | `0.75` | 0 | 1 | - | How much to transform the reference image. 0 means no change, 1 means complete transformation | | `controlImage` | file\_array | No | - | - | - | - | Control images to use as input (edge, depth, or pose). Cannot be used with Reference Image. | | `controlnetConditioningScale` | number | No | `0.7` | 0 | 1 | - | The scale of the controlnet conditioning | | `numOutputs` | number | No | `4` | 1 | 4 | - | Number of images to generate | | `numInferenceSteps` | number | No | `9` | 1 | 100 | - | Number of inference steps. | | `width` | number | No | - | 128 | 2048 | - | Width of the generated image | | `height` | number | No | - | 128 | 2048 | - | Height of the generated image | | `seed` | number | No | - | 0 | 2147483647 | - | Used to reproduce previous results |