Models and generation settings

The catalog below is generated from the same configuration as Create and Studio. Prices are credits per generation at the listed default settings, not dollars or per-second rates.

Gemini Omni Flash guide: upload an image or video and describe your changes. Video output, 720p, flat price per clip up to 10 seconds.

Background image generation

Image jobs can return a running status. Poll the job until completed or use a completion webhook, then read output.imageUrl. Keep the job ID when a request is interrupted; submitting again creates a separate job.

Nano Banana Pro also supports an optional batch mode when enabled. Batch can take up to 24 hours and occasionally longer. Use it for queued work rather than interactive previews. Credits currently match standard generation; failed jobs receive the normal refund.

{
  "workflow": "generate_image",
  "input": {
    "model": "nano-banana-pro",
    "prompt": "A ceramic vase in soft studio lighting",
    "resolution": "1K",
    "aspect_ratio": "1:1",
    "execution_mode": "batch"
  }
}

Send this body to /api/v1/jobs. Omit execution_mode for standard generation. Batch currently supports Nano Banana Pro only and returns an error before charging if it is not enabled.

Model / catalog ID	Workflow	Default credits	Settings
Seedance 2.5 Video Edit `seedance-2.5-edit`	`generate_video`	3070	480p / 720p · Reference image supported · Video reference supported Defaults: {"resolution":"720p"}
Seedance 2.5 Video Reference `seedance-2.5-ref`	`generate_video`	2730	480p / 720p · 5 / 10 / 15 / 20 / 30 seconds · Reference image supported · Video reference supported Defaults: {"duration":5,"resolution":"720p"}
Gemini Omni 1.1 Flash `gemini-omni-flash`	`generate_video`	413	720p · 10 seconds · Reference image supported · Video reference supported Defaults: {"duration":10,"resolution":"720p"}
Nano Banana `nano-banana`	`generate_image`	20	1K Defaults: {"resolution":"1K"}
Nano Banana 2 Lite `nano-banana-2-lite`	`generate_image`	20	1K Defaults: {"resolution":"1K"}
Nano Banana 2 `nano-banana-2`	`generate_image`	35	1K / 2K / 4K Defaults: {"resolution":"1K"}
Nano Banana Pro `nano-banana-pro`	`generate_image`	75	1K / 2K / 4K · Reference image supported Defaults: {"resolution":"1K"}
Talking portrait `lipsync` Studio node	`lipsync_video`	100	Reference image supported Defaults: {}
Motion transfer `motion-control` Studio node	`transfer_motion`	750	Reference image supported · Video reference supported Defaults: {}
ElevenLabs Speech `elevenlabs-tts`	`generate_speech`	1	See workflow guide Defaults: {}
Prompt assistant `prompt-assistant`	`improve_prompt`	3	See workflow guide Defaults: {}
Gemini Image `gemini-image`	`generate_image`	20	See workflow guide Defaults: {}
GPT Image 2 `gpt-image-2`	`generate_image`	30	See workflow guide Defaults: {}
GPT Image 2.5 Sunburst `gpt-image-2.5-sunburst`	`generate_image`	15	Reference image supported Defaults: {}
GPT Image 2.5 Flare `gpt-image-2.5-flare`	`generate_image`	15	Reference image supported Defaults: {}
Grok Image `grok-image`	`generate_image`	8	See workflow guide Defaults: {"model":"grok-imagine"}
Gemini Image Edit `gemini-edit`	`edit_image`	20	Reference image supported Defaults: {}
Seedance 2.5 `seedance-2.5`	`generate_video`	500	480p / 720p / 1080p · 5 / 10 / 15 / 20 / 30 seconds · Reference image supported Defaults: {"duration":5,"resolution":"720p"}
Seedance 2 `seedance-2`	`generate_video`	400	480p / 720p · 5 / 10 / 15 seconds · Reference image supported Defaults: {"duration":5,"resolution":"720p"}
Kling `kling-3.1`	`generate_video`	250	5 / 10 seconds · Reference image supported Defaults: {"duration":5}
Veo 3 `veo3`	`generate_video`	110	Reference image supported Defaults: {}
Grok Video `grok-video`	`generate_video`	150	5 / 6 / 10 / 15 seconds · Reference image supported Defaults: {"model":"grok-imagine","duration":6}

GET /api/v1/creator/models returns these definitions under data, including ratios, durations, resolutions, reference, extraInputs, studioOnly, and estimatedCredits. It requires authentication and jobs:read. An entry describes the implementation; successful generation also depends on the deployment having the corresponding provider configured.

Nano Banana and GPT Image

The table above is generated from the same pricing catalog used by Create and the API. Nano Banana variants expose text-to-image; use the editing workflow for reference edits. GPT Image outputs support 1024×1024, 1536×1024, or 1024×1536. GPT Image 2.5 supports medium/high quality and up to five ordered reference images. Medium is 15 credits, high 30, plus 10 per reference. Image prompts are limited to 4,000 UTF-8 bytes. Talking portraits are temporarily unavailable pending provider cost verification.

POST /api/v1/jobs · Nano Banana Pro, 4Kjson

{
  "workflow": "generate_image",
  "input": {
    "resolution": "4K",
    "prompt": "A sun oracle in intricate gold armor, photoreal portrait",
    "aspect_ratio": "3:4",
    "model": "nano-banana-pro"
  }
}

Duration and batch pricing

Video pricing scales with duration and resolution. Check the current catalog quote before submitting. Four variations create four independently charged jobs. There is no batch parameter on the canonical job endpoint; submit one job per variation.

Speech costs one credit per started 100 characters, and prompt refinement costs one credit per request. An empty-script catalog estimate is not the final speech charge. Studio estimates depend on graph inputs and may differ from actual usage when text is produced by an earlier node.

Catalog IDs describe workspace nodes. For example, the GPT Image 2 entry maps to the gpt-image-2 job model, while speech uses generate_speech and voice_id. Use Create’s “View API request” or the specific image, video, and speech guides. See billing to convert your purchased credits to a dollar cost.