Skip to main content
Pass a public model ID in the top-level model field. GENGEN uses that ID to select the provider adapter, validate supported controls, and route the request.

​
Browse the catalog

The model pages are generated from the same catalog used by the GENGEN application, so public model IDs, supported modes, inputs, outputs, and endpoint assignments stay aligned with the runtime configuration.

Video models

Video generation, animation, and enhancement models.

Image models

Synchronous and asynchronous image generation and editing.

LLM

Chat completions, multimodal reasoning, and video understanding.

​
OpenAI

GPT Image 2.5 Flare, GPT Image 2.5 Sunburst, and GPT Image 2 support synchronous generation and reference-image editing through the OpenAI Image API. Flare emphasizes speed for everyday generation; Sunburst emphasizes editing precision. All three use GENGEN’s normalized request fields and return image URLs retained for 24 hours. GPT Image 2.5 adds xhigh and max quality; streaming is not exposed by this endpoint.

​
LLM and video understanding

Gemini 3.8 Flash supports text generation and multimodal understanding. Use Chat Completions for synchronous or streaming text and Responses for video analysis.

​
Video providers

Kling 3.0 Turbo

Fast Kling video generation for text-to-video and first-frame image-to-video workflows.

Kling 3.0

Kuaishou video generation with native audio, multi-shot control, and output up to 4K.

Kling 3.0 Omni

Multimodal video generation from text, images, elements, and reference or base video.

MiniMax H3

Text, first-and-last-frame, and multimodal reference video generation up to 2K.

MiniMax H3 Max

Fast text-to-video and image-to-video generation at 480P or 768P.
Capabilities, availability, and final customer prices can differ by model. Check the Models section in the Dashboard before production use.

​
Model selection

  • Use the exact public ID shown by GENGEN.
  • Do not send upstream provider model aliases unless a compatibility page explicitly documents them.
  • Keep provider-specific overrides under providerOptions.
  • Keep media under assets and provider-neutral settings under controls.

​
Source of truth

The Dashboard model catalog remains the source of truth for current availability and final customer prices. The documentation does not publish internal provider pricing formulas.

​
Higgsfield

Seedance 2.0 (Higgsfield) and Seedance 2.5 (Higgsfield) are independent provider entries. Use asynchronous generation with wallet billing. See the Higgsfield guide.