Skip to main content
    SeamUI LogoSeamUI
    Product
    PricingFAQModelsGuides
    Sign inStart free trial

    AI MODELS

    31 AI models.
    Images and videos. Your own keys.

    Every model below runs on the API key you paste into SeamUI, from OpenAI, Google, Black Forest Labs, xAI, Google DeepMind, Byteplus and Kuaishou. You pay the provider directly, we never sit between you and your usage.

    15 image models and 16 video models are ready today.

    Explore video models

    Three ways to reach them

    Pick a route per generation. Aggregators cover several families with one key, direct keys keep the request between you and the vendor.

    Direct provider

    One key per vendor, straight to OpenAI, Google, Black Forest Labs or xAI. No middleman on the request.

    15 image models, 12 video models

    Kie.ai

    One aggregator key unlocks several families at once, with 1K to 4K resolution tiers on the top models.

    12 image models, 14 video models

    Get a key

    The catalog

    Filter by output, model family or the route you plan to use.

    OUTPUT
    FAMILY
    ROUTE

    31 models

    GPT Image 2.5 Flare

    OpenAI

    OpenAI's fastest 2.5 tier, for high quality everyday generation at volume.

    IMAGE
    • ·Fastest GPT Image tier
    • ·Quality up to max
    • ·Up to 16 reference images
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 16 references
    Aspect ratios9 presets
    Resolution tiers1K / 2K / 4K
    Max prompt32k characters
    Vendor docs

    GPT Image 2.5 Sunburst

    OpenAI

    OpenAI's precision editing tier, built for edits that land exactly where you ask.

    IMAGE
    • ·Precise local edits and inpainting
    • ·Quality up to max
    • ·Transparent backgrounds
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 16 references
    Aspect ratios9 presets
    Resolution tiers1K / 2K / 4K
    Max prompt32k characters
    Vendor docs

    GPT Image 2

    OpenAI

    OpenAI's flagship image model, strongest at legible text and precise instruction following.

    IMAGE
    • ·Reliable text rendering
    • ·Pixel-exact output sizes
    • ·Up to 16 reference images
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 16 references
    Aspect ratios5 presets
    Resolution tiers1K / 2K / 4K
    Max prompt32k characters
    Vendor docs

    GPT Image 1.5

    OpenAI

    OpenAI's previous flagship image model, with strong prompt adherence and precise image editing.

    IMAGE
    • ·Strong instruction following
    • ·Transparent backgrounds
    • ·Up to 16 reference images
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 16 references
    Aspect ratios3 presets
    Max prompt32k characters
    Vendor docs

    GPT Image 1

    OpenAI

    OpenAI's original GPT Image model for generation, editing, and high-fidelity reference workflows.

    IMAGE
    • ·Image generation and editing
    • ·Input fidelity control
    • ·Transparent backgrounds
    Available onDirect provider
    Text to imageYes
    Image to imageUp to 16 references
    Aspect ratiosFixed sizes
    Max prompt32k characters
    Vendor docs

    GPT Image 1 Mini

    OpenAI

    OpenAI's cost-efficient GPT Image model for high-volume generation and editing.

    IMAGE
    • ·Lowest GPT Image cost
    • ·Image generation and editing
    • ·Transparent backgrounds
    Available onDirect provider
    Text to imageYes
    Image to imageUp to 16 references
    Aspect ratiosFixed sizes
    Max prompt32k characters
    Vendor docs

    Nano Banana 2 (Gemini 3.1 Flash)

    Google

    The fast Nano Banana, the best default for high-volume batches.

    IMAGE
    • ·Fast turnaround
    • ·Widest aspect ratio range
    • ·Up to 14 reference images
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 14 references
    Aspect ratios14 presets
    Resolution tiers1K / 2K / 4K / 512
    Max prompt32k characters
    Vendor docs

    Nano Banana Pro (Gemini 3 Pro)

    Google

    Google's top-tier image model, built for composition-heavy scenes and careful edits.

    IMAGE
    • ·High prompt fidelity
    • ·Up to 14 reference images
    • ·1K to 4K on kie.ai
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 14 references
    Aspect ratios10 presets
    Resolution tiers1K / 2K / 4K
    Max prompt32k characters
    Vendor docs

    Nano Banana (Gemini 2.5 Flash)

    Google

    The original Nano Banana, still a cheap workhorse for simple edits.

    IMAGE
    • ·Low cost per image
    • ·Solid character consistency
    • ·Broad provider coverage
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 10 references
    Aspect ratios10 presets
    Max prompt32k characters
    Vendor docs

    Flux 2 Pro

    Black Forest Labs

    Photoreal generation with a seed control, so a look can be reproduced run after run.

    IMAGE
    • ·Photorealism
    • ·Seed and safety tolerance controls
    • ·Up to 8 reference images
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 8 references
    Aspect ratios7 presets
    Resolution tiers1K / 2K
    Max prompt4k characters
    Vendor docs

    Flux 2 Max

    Black Forest Labs

    The largest Flux 2 tier, for the shots where detail matters more than latency.

    IMAGE
    • ·Highest Flux detail
    • ·Seed control
    • ·Up to 8 reference images
    Available onDirect provider
    Text to imageYes
    Image to imageUp to 8 references
    Aspect ratiosFixed sizes
    Max prompt4k characters
    Vendor docs

    Flux Kontext Pro

    Black Forest Labs

    Edit-first Flux, made to change one thing in an image and leave the rest alone.

    IMAGE
    • ·In-place editing
    • ·Keeps the untouched areas stable
    • ·One reference on Kie
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 4 references
    Aspect ratios10 presets
    Max prompt4k characters
    Vendor docs

    Flux Kontext Max

    Black Forest Labs

    The strongest Kontext tier for demanding, multi-step edits.

    IMAGE
    • ·Hardest edits
    • ·Strong typography retention
    • ·One reference on Kie
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 4 references
    Aspect ratios10 presets
    Max prompt4k characters
    Vendor docs

    Grok Imagine Image 2.0

    xAI

    xAI's quality-tier image model for production-ready generation and precise natural-language edits.

    IMAGE
    • ·Higher realism and text rendering
    • ·Up to 5 references via Kie
    • ·Direct 1K and 2K output
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 5 references
    Aspect ratios16 presets
    Resolution tiers1K / 2K
    Max prompt1.024k characters
    Vendor docs

    Grok Imagine Image

    xAI

    xAI's image model, fast and loose on style, useful for concepting.

    IMAGE
    • ·Quick concepting
    • ·6 images per generation on Kie
    • ·Available on all three routes
    Available onDirect provider, Kie.ai
    Text to imageYes
    Image to imageUp to 3 references
    Aspect ratios15 presets
    Resolution tiers1K / 2K
    Max prompt4k characters
    Vendor docs

    Veo 3.1

    Google DeepMind

    Preview

    Google's flagship video model, the quality tier with native audio.

    VIDEO
    • ·Native audio
    • ·Text and image to video
    • ·Up to 4K
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame, Reference images, Source video
    Kie.ai inputsFirst frame, Last frame
    Other capabilitiesClip extension
    Max prompt5k characters
    Vendor docs

    Veo 3.1 Fast

    Google DeepMind

    Preview

    The fast Veo tier, and the cheapest way to iterate on a shot.

    VIDEO
    • ·Fast turnaround
    • ·Material reference mode
    • ·Up to 3 reference images
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame, Reference images, Source video
    Kie.ai inputsFirst frame, Last frame, Reference images
    Other capabilitiesClip extension
    Max prompt5k characters
    Vendor docs

    Veo 3.1 Lite

    Google DeepMind

    Preview

    The lightest Veo tier, for drafts and volume.

    VIDEO
    • ·Lowest Veo cost
    • ·Material reference mode
    • ·Text and image to video
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame
    Kie.ai inputsFirst frame, Last frame, Reference images
    Other capabilitiesModel dependent
    Max prompt5k characters
    Vendor docs
    B

    Seedance 2.5

    Byteplus

    The newest Seedance, built for long takes and multimodal references.

    VIDEO
    • ·Up to 30 seconds
    • ·Up to 1080p
    • ·Up to 30 reference images
    • ·mp4 or mov output
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Kie.ai inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Other capabilitiesNative audio
    Max prompt20k characters
    Vendor docs
    B

    Seedance 2.0

    Byteplus

    Preview

    Bytedance's multimodal video model, strong on multi-shot consistency.

    VIDEO
    • ·Image, video and audio references
    • ·Native audio
    • ·4 to 15 seconds
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Kie.ai inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Other capabilitiesNative audio
    Max prompt20k characters
    Vendor docs
    B

    Seedance 2.0 Fast

    Byteplus

    Preview

    The fast Seedance tier, capped at 720p.

    VIDEO
    • ·Fast turnaround
    • ·Multimodal references
    • ·Native audio
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Kie.ai inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Other capabilitiesNative audio
    Max prompt20k characters
    Vendor docs
    B

    Seedance 2.0 Mini

    Byteplus

    The cheapest Seedance tier, same multimodal contract as 2.0.

    VIDEO
    • ·Lowest Seedance cost
    • ·Multimodal references
    • ·Native audio
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Kie.ai inputsFirst frame, Last frame, Reference images, Reference video, Reference audio
    Other capabilitiesNative audio
    Max prompt20k characters
    Vendor docs
    B

    Seedance 1.5 Pro

    Byteplus

    Preview

    The previous Seedance generation, with a fixed-lens option for static shots.

    VIDEO
    • ·Fixed lens control
    • ·Frame pair or single image
    • ·4 to 12 seconds
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Last frame
    Kie.ai inputsReference images
    Other capabilitiesNative audio
    Max prompt20k characters
    Vendor docs
    B

    Seedance 1.0 Pro

    Byteplus

    The 1.0 generation, silent output and the widest duration floor at 2 seconds.

    VIDEO
    • ·From 2 seconds
    • ·Frame pair support
    • ·Seed control
    Available onDirect provider
    Text to videoYes
    Direct provider inputsFirst frame, Last frame
    Other capabilitiesModel dependent
    Max prompt20k characters
    Vendor docs
    B

    Seedance 1.0 Pro Fast

    Byteplus

    The fast 1.0 tier, first frame or text only.

    VIDEO
    • ·Lowest cost per second
    • ·From 2 seconds
    • ·Seed control
    Available onDirect provider
    Text to videoYes
    Direct provider inputsFirst frame
    Other capabilitiesModel dependent
    Max prompt20k characters
    Vendor docs

    Grok Imagine Video 1.5

    xAI

    Preview

    xAI's newest video model, and the one it recommends for video.

    VIDEO
    • ·Native 1080p
    • ·1 to 15 seconds
    • ·Image, reference and video inputs
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Reference images, Last frame
    Kie.ai inputsReference images
    Other capabilitiesModel dependent
    Max prompt5k characters
    Vendor docs

    Grok Imagine Video (1.0)

    xAI

    Preview

    xAI's video model, the only one that can extend a clip it already made.

    VIDEO
    • ·6 to 30 seconds
    • ·Extends its own videos
    • ·Up to 7 reference images
    Available onDirect provider, Kie.ai
    Text to videoYes
    Direct provider inputsFirst frame, Reference images, Reference video, Reference video
    Kie.ai inputsReference images, Source video
    Other capabilitiesClip extension
    Max prompt5k characters
    Vendor docs
    K

    Kling 3.0

    Kuaishou

    Kuaishou's flagship video model, from a prompt or from a frame pair.

    VIDEO
    • ·3 to 15 seconds
    • ·720p to 4K
    • ·Native audio
    Available onKie.ai
    Text to videoYes
    Kie.ai inputsFirst frame, Last frame, Reference images
    Other capabilitiesNative audio
    Max prompt2.5k characters
    Vendor docs
    K

    Kling 3.0 Turbo

    Kuaishou

    The fast Kling tier, for drafts and iterations.

    VIDEO
    • ·3 to 15 seconds
    • ·720p or 1080p
    • ·Fastest Kling tier
    Available onKie.ai
    Text to videoYes
    Kie.ai inputsFirst frame
    Other capabilitiesModel dependent
    Max prompt2.5k characters
    Vendor docs
    K

    Kling 3.0 Omni

    Kuaishou

    The complete Kling contract: references, video editing and native audio.

    VIDEO
    • ·Up to 7 reference images
    • ·Imitates or edits a video
    • ·3 to 15 seconds
    Available onKie.ai
    Text to videoYes
    Kie.ai inputsFirst frame, Last frame, Reference images, Reference video, Reference video
    Other capabilitiesNative audio
    Max prompt2.5k characters
    Vendor docs
    K

    Kling 3.0 Motion Control

    Kuaishou

    Transfers the motion of a source video onto a character image.

    VIDEO
    • ·Character image plus source video
    • ·720p or 1080p
    • ·Follows the image or the video
    Available onKie.ai
    Text to videoYes
    Kie.ai inputsFirst frame, Reference video
    Other capabilitiesModel dependent
    Max prompt2.5k characters
    Vendor docs

    Retired models

    These ids are no longer generated. Old prompts and presets still open: SeamUI migrates them to the replacement model automatically.

    • DALL·E 3 · OpenAInow GPT Image 2
    • DALL·E 2 · OpenAInow GPT Image 2
    • Grok Imagine Pro · xAInow Grok Imagine Image
    • Gemini Omni Flash · Googlenow Veo 3.1 Fast

    Questions

    Which AI video models does SeamUI support?
    SeamUI supports production-ready models across Veo 3.1, Seedance and Grok Imagine Video. The exact models and routes shown here come from the same registry used by the app.
    Can I generate video from text, an image or an existing clip?
    Yes, depending on the model. SeamUI shows each model's supported inputs, including first and last frames, reference images, reference video, reference audio and source-video editing where available.
    Does SeamUI resell AI image or video models?
    No. You bring your own API keys and pay each provider directly for what you generate. SeamUI is the workspace that drives those keys at scale, it never marks up your usage.
    What is the difference between a direct key and an aggregator key?
    A direct key talks to one vendor: OpenAI, Google, Black Forest Labs or xAI. An aggregator key from Kie.ai covers several model families at once, which is the fastest way to start if you only want to manage a single key.
    What happens to my presets when a model is retired?
    Retired model ids stay readable. When you reuse an old prompt or preset, SeamUI migrates it to the current replacement model and normalizes the settings that model actually supports, so nothing in your history breaks.
    Can I use several models in the same project?
    Yes. Model and provider are picked per generation, so you can run the same prompt across models and compare the results inside one project.

    Create the whole campaign,
    not one asset at a time.

    Start your 7-day free trial

    Product

    • The workspace
    • Pricing
    • FAQ
    • Image & video models
    • Prompt inspirations
    • Free AI image generator
    • Free AI video generator
    • Kie.ai image & video generation
    • Alternatives & comparisons
    • Partners
    • Affiliate program

    Use cases

    • For agencies
    • For Shopify sellers
    • For Etsy sellers
    • For Amazon sellers
    • For course creators
    • For Instagram creators
    • For LinkedIn creators
    • For real estate
    • For TikTok creators

    Guides

    • ChatGPT guides
    • Gemini guides
    • Grok guides
    • AI video guides
    • AI workflow guides
    • All guides

    Popular questions

    • How do I batch generate images with ChatGPT?
    • How many images can ChatGPT generate per day?
    • How do I use reference images in ChatGPT?
    • Can I use ChatGPT images commercially?
    • Can ChatGPT generate images from a spreadsheet?

    SeamUI © 2026 SeamUI. Pay once, create forever.

    Sign inPrivacyTermsContactSitemap
    SeamUI