PricingAPI keys
OverviewQuickstartAuthentication
GenerateModelsImagesVideoUploads
JobsJobs and pollingWebhooks
ReferenceErrors and retriesLimitsClient
BillingBilling
Media toolsOverviewAvatarRemix mediaFace swapMusicTimelineVideo captions
MCPOverviewQuickstartOAuthTools and gates
CLIOverviewInstall and loginCommandsAgent skills
API referenceAPI reference
EnterpriseOverviewAgentsFormatsBest practices

Generate

Models

Pick the model yourself: prices differ by more than a hundred times between models. Image models run on the Image Router; video, lip-sync and captions models on the Video Router.

List models

GEThttps://api.sume.com/v1/image-router/models

GEThttps://api.sume.com/v1/video-router/models

Both need a key and return the models your workspace can run right now, with their capabilities and prices. Add /{model_id} for one model.

cURL
curl https://api.sume.com/v1/image-router/models/gpt-image-2 \
  -H "Authorization: Bearer $SUME_API_KEY"

A model row

capabilities says which fields and values the model takes; pricing gives its price and its basis (per_image, per_second or per_second_by_resolution); constraints lists the rules in plain words.

200 OK
{
  "data": {
    "id": "gpt-image-2",
    "name": "ChatGPT Image 2",
    "modality": "image",
    "capabilities": {
      "text_to_image": true,
      "image_edit": true,
      "aspect_ratios": [
        "auto",
        "1:1",
        "16:9",
        "9:16",
        "4:3",
        "3:4"
      ],
      "qualities": [
        "low",
        "medium",
        "high"
      ],
      "max_images": 4,
      "output_formats": [
        "png",
        "jpeg",
        "webp"
      ]
    },
    "pricing": {
      "basis": "per_image"
    },
    "constraints": [
      "…"
    ]
  }
}

Models on Sume

Pass the id as model to the endpoint shown. Each name links to the model's page with examples. The list endpoints above are the live source; a model can be added before this table is.

Image

Model idModelWhat it doesEndpoint
gpt-image-2.5GPT Image 2.5 FlareGenerates and edits images · up to 4 per request/v1/image-router/generate
gpt-image-2.5-sunburstGPT Image 2.5 SunburstGenerates and edits images · up to 4 per request/v1/image-router/generate
gpt-image-2ChatGPT Image 2Generates and edits images · up to 4 per request/v1/image-router/generate
nano-banana-2.1Nano Banana 2.1Generates and edits images · up to 4K · up to 4 per request/v1/image-router/generate
nano-banana-proNano Banana ProGenerates and edits images · up to 4K · up to 4 per request/v1/image-router/generate
seedream-5-liteSeedream 5.0 LiteGenerates and edits images · up to 4 per request/v1/image-router/generate
seedream-4-5-editSeedream 4.5Generates and edits images · up to 4 per request/v1/image-router/generate
grok-imageGrok ImagineGenerates and edits images/v1/image-router/generate
seedream-v4Seedream 4.0Generates and edits images · up to 4 per request/v1/image-router/generate
ideogram-v3Ideogram V3Generates and edits images · up to 4 per request/v1/image-router/generate
recraft-v4Recraft V4Generates images from text · up to 4 per request/v1/image-router/generate
qwen-image-layeredQwen Image LayeredEdits images · up to 4 per request/v1/image-router/generate
ideogram-v4.5Ideogram 4.5Generates and edits images · up to 2K · up to 4 per request/v1/image-router/generate

Video

Model idModelWhat it doesEndpoint
seedance-2.5Seedance 2.5Generates video from text, an image, references · 480p to 1080p · 4–30 s · native audio/v1/video-router/generate
seedance-2-miniSeedance 2.0 MiniGenerates video from text, an image, references · 480p to 720p · 4–15 s · native audio/v1/video-router/generate
seedance-2Seedance 2.0Generates video from text, an image, references · 480p to 1080p · 4–15 s · native audio/v1/video-router/generate
seedance-2-fastSeedance 2.0 FastGenerates video from text, an image, references · 480p to 720p · 4–15 s · native audio/v1/video-router/generate
kling-4Kling 4.0Generates video from text, an image · 720p to 1080p · 3–30 s · native audio/v1/video-router/generate
kling-4-omniKling 4.0 OmniGenerates video from text, an image, references · 720p to 1080p · 3–30 s · native audio/v1/video-router/generate
kling-3Kling Video v3 ProGenerates video from text, an image · 1080p · 4–15 s · native audio/v1/video-router/generate
wan-3.0Wan 3.0Generates video from text, an image, references · 480p to 1080p · 2–30 s · native audio/v1/video-router/generate
grok-imagine-video-1.5Grok Imagine Video 1.5Generates video from an image · 480p to 720p · 4–15 s/v1/video-router/generate
minimax-h3MiniMax H3Generates video from text, an image, references · 768p to 4K · 5–15 s · native audio/v1/video-router/generate
minimax-h3-maxMiniMax H3 MaxGenerates video from text, an image, references · 480p to 1080p · 5–15 s · native audio/v1/video-router/generate
gemini-omni-flash-1.1Gemini Omni Flash 1.1Generates video from text, an image, references, a source video · 360p to 4K · 3–10 s · native audio/v1/video-router/generate
veo-3.1Veo 3.1Generates video from text · 720p · 4–8 s · native audio/v1/video-router/generate
veo-3.1-fastVeo 3.1 FastGenerates video from text · 720p · 4–8 s · native audio/v1/video-router/generate
veo-3.1-liteVeo 3.1 LiteGenerates video from text · 720p · 4–8 s · native audio/v1/video-router/generate
wand-vega-1.0-liteWAND-Vega 1.0 LiteGenerates video from text · 768p to 4K · 4–15 s · native audio/v1/video-router/generate
wand-vega-1.0-proWAND-Vega 1.0 ProGenerates video from text · 720p to 4K · 4–15 s · native audio/v1/video-router/generate
wand-vega-1.5-proWAND-Vega 1.5 ProGenerates video from text · 720p to 4K · 4–15 s · native audio/v1/video-router/generate
mingmou-1Mingmou 1Generates video from text · 720p to 4K · 5–10 s/v1/video-router/generate
hailuo-h3-regenHailuo H3 RegenGenerates video from a source video · 2K · 5–16 s/v1/video-router/generate
heygen-video-1HeyGen Video 1Generates video from text, an image, references · 480p to 2K · 5–30 s · native audio/v1/video-router/generate

Lip-sync

Model idModelWhat it doesEndpoint
veed-fabric-1.0VEED Fabric 1.0Makes a face say an audio track · 480p to 720p/v1/video-router/generate
kling-avatar-2.0-proKling Avatar 2.0 ProMakes a face say an audio track/v1/video-router/generate
kling-avatar-2.0Kling Avatar 2.0Makes a face say an audio track/v1/video-router/generate
infinitalkInfiniteTalkMakes a face say an audio track · 480p to 720p/v1/video-router/generate
omnihuman-1.5OmniHuman 1.5Makes a face say an audio track · 720p to 1080p/v1/video-router/generate
sync-lipsync-3Sync Lipsync 3Makes a face say an audio track/v1/video-router/generate
sync-lipsync-2-proSync Lipsync 2 ProMakes a face say an audio track/v1/video-router/generate
sync-lipsync-2Sync Lipsync 2Makes a face say an audio track/v1/video-router/generate
kling-lipsyncKling LipsyncMakes a face say an audio track/v1/video-router/generate
latentsyncLatentSyncMakes a face say an audio track/v1/video-router/generate
veed-lipsyncVEED LipsyncMakes a face say an audio track/v1/video-router/generate
infinitalk-videoInfiniteTalk VideoMakes a face say an audio track · 480p to 720p/v1/video-router/generate
pixverse-lipsyncPixVerse LipsyncMakes a face say an audio track/v1/video-router/generate

Captions

Model idModelWhat it doesEndpoint
veed-subtitlesVEED SubtitlesBurns styled captions into a video · 1080p/v1/video-router/generate

Related

  • Images
  • Video
  • Limits and billing

On this page

  1. List models
  2. A model row
  3. Models on Sume
  4. Image
  5. Video
  6. Lip-sync
  7. Captions
  8. Related