First month 40% off — unlock Seedance 2.5, MiniMax H3 and more models for $10.99First month 40% off — Seedance 2.5, MiniMax H3 & more for $10.99

Top AI Video & Image Models

Try Seedance 2.5, Veo 3.1, Kling 3.0, GPT Image 2.5, Nano Banana 2 and more in one place. Compare specs side by side, then create ads, social clips and product videos.

Video Models

ModelAudioReferenceDurationResolution
Wizshot 1.0NewNative audioOmni5-15s720PStart Creating
Seedance 2.5Native audioOmni4-30s1080pStart Creating
Seedance 2.0Native audioOmni4-15s4KStart Creating
Seedance 1.5SilentImage4-12s1080pStart Creating
MiniMax H3Native audioOmni4-15s2KStart Creating
Kling 4.0NewOmni3-30s2KStart Creating
Kling 3.0 OmniNative audioImage4-15s4KStart Creating
Gemini Omni 1.1 FlashNewNative audioImage · Video3-10s4KStart Creating
Gemini Omni FlashNative audioImage · Video3-10s1080pStart Creating
Veo 3.1Native audioImage4-8s4KStart Creating

Image Models

ModelResolutionCredits
GPT Image 2.5New4K5+Start Creating
GPT Image 24K3+Start Creating
Nano Banana 24K6+Start Creating
Seedream 5.02K6+Start Creating

Which model should I choose?

Frequently asked questions

  • Which model should I choose for the best video quality?

    Seedance 2.5 is the flagship for overall quality and native audio. For synced audio with strong prompt adherence, Veo 3.1 is a great alternative.

  • What is the difference between video and image models?

    Video models generate motion clips (with or without audio); image models generate still images. Use the category tabs above to compare each group.

  • Do the models support image references?

    Most models accept text and image references. The reference column in each table shows what every model supports.

  • Why do the available models change over time?

    The model lineup and parameters follow the live creation workbench. This page is a snapshot; always check the workbench for the latest availability.

  • Can I switch models without losing my prompt?

    Yes. In the workbench you can switch models and keep your prompt and settings to compare outputs quickly.

  • Is there a cost difference between models?

    Different models consume different amounts of credits. The workbench shows the exact cost before you generate.

  • Which models can use a video as a reference?

    Seedance 2.5, Seedance 2.0, Wizshot 1.0 and MiniMax H3 support Omni references, so you can mix image, video and audio references. Gemini Omni 1.1 Flash and Gemini Omni Flash accept image and video references, while Kling 3.0 Omni, Veo 3.1 and Seedance 1.5 accept image references.

Compare the Models and Start Creating

Pick the model that fits your shot, then open the workbench and generate in minutes.