Models

Browse every model with its prices at geekidea.ai/models, or list them with the API. GET /v1/models and GET /v1/models/{author}/{model} work without an API key; with a key, prices are your organization’s actual prices.

curl https://api.geekidea.ai/v1/models?type=image

Model IDs

Model IDs are author/model, where the author is the company that made the model, for example google/nano-banana-pro or bytedance/seedance-2.5. Some models also have aliases (listed in aliases), which you can use anywhere a model ID is accepted.

The model object

{
  "id": "midjourney/midjourney-v8.2",
  "object": "model",
  "created": 1791313267,
  "owned_by": "midjourney",
  "name": "Midjourney V8.2",
  "type": "image",
  "aliases": ["mj-v8.2"],
  "modalities": { "input": ["text", "image"], "output": ["image"] },
  "capabilities": ["text_to_image", "image_to_image", "reference_image"],
  "context_length": null,
  "pricing": { "currency": "USD", "image": 0.092, "image_per_spec": { "1K": 0.092, "2K": 0.138 } },
  "generation_config": {
    "n": { "type": "integer", "default": 1, "max": 2 },
    "resolution": { "type": "string", "enum": ["1K", "2K"] },
    "aspect_ratio": { "type": "string", "enum": ["16:9", "3:2", "4:3", "1:1", "3:4", "2:3", "9:16"], "default": "1:1" },
    "images_per_request": { "type": "integer", "value": 4 },
    "max_reference_images": { "type": "integer", "max": 2 }
  }
}

Field	Meaning
`type`	`text`, `image` or `video`. Filter the list with `?type=`
`modalities`	Accepted inputs and produced outputs
`capabilities`	What the model can do, for example `image_to_image`, `first_frame`, `reference_video`, `generate_audio`
`context_length`	Context window of text models, when known
`pricing`	Prices in US dollars, see below
`generation_config`	The parameters this model accepts and their allowed values

Pricing fields

Only the fields that apply to a model are present.

Field	Unit
`input`, `cached_input`, `output`	Per million tokens. For token-billed image models, `output` is per million output image tokens
`image`	Per image (per request for models such as Midjourney that return several images per request). Also the price of any quality or resolution tier not listed in `image_per_spec`
`image_per_spec`	Per image, by quality or resolution tier (`low`, `high`, `1K`, `2K`…)
`second`	Per second of video
`second_per_spec`	Per second, by `resolution\|audio` or `resolution\|silent`, for example `720p\|audio`. For token-billed video models (Seedance) this is a reference price at the typical output size
`video_tokens`, `video_tokens_per_spec`	Token-billed video models: per million video tokens, by `resolution\|audio`
`video_tokens_video_input_per_spec`	Token-billed video models: per million video tokens when reference videos are supplied

See Billing for how calls are charged.

generation config

Each entry in generation_config describes one parameter:

Key	Meaning
`type`	`string`, `integer` or `boolean`
`enum`	Allowed values
`min`, `max`	Allowed range
`default`	Used when you leave the parameter out
`value`	A fixed behavior of the model, not something you send, for example `images_per_request`

Where each parameter goes in the request depends on the endpoint. For images, resolution and aspect_ratio go in the request’s generation_config object; for videos, duration, resolution and aspect_ratio are top-level fields. Every model page on geekidea.ai/models shows the exact field path. Values outside the allowed set are rejected with 400. Input limits such as max_reference_images, max_reference_videos and max_reference_audios cap how many references you can send.

Get started

Generate

Workflows

Async and webhooks

Account

Model IDs

The model object

Pricing fields

generation config

​Model IDs

​The model object

​Pricing fields

​generation config

Model IDs

The model object

Pricing fields

generation config