AI video models

Seedance 2.0 / Capabilities — Text accepts no media; frames requires a first frame and allows a last frame; reference requires a prompt and at least one image, video, or audio reference. Reference media cannot be sent with endpoint-frame fields. Seedance 2.0 supports three mutually exclusive text, frame, and multimodal-reference scenes. The current callable output set is 720p, 1080p, and 4K for 4–15 seconds; video-reference requests are available only at 1080p or 4K because those are the current priced combinations. 164–3,840 credits per request

Seedance 2.0 Fast / Capabilities — Text accepts no media; frames requires a first frame and allows a last frame; reference requires a prompt and at least one image, video, or audio reference. Reference media cannot be sent with endpoint-frame fields. 480p/720p · 4–15 · 47–450 credits per request.

Seedance 1.5 Pro / Capabilities — Seedance 1.5 Pro on Kyeo AI accepts text alone or up to two images and exposes 480p/720p/1080p, 4–12 seconds, generated audio, six ratios, and a fixed-lens switch. The 7–180-credit cost follows the selected combination; the Pro name and higher credit cost do not prove image quality, stability, or publication readiness.

HappyHorse 1.0 / Workflows — The currently available workflows are prompt-only video, one-image video, and reference video with 2 to 9 images. Editing an existing video is currently unavailable on this page.

Wan 2.7 supports text-to-video, first-frame or first-and-last-frame image-to-video, and video editing. It currently offers 720p and 1080p, so compare media requirements, duration, framing, and estimated credits together.

No media selects text-to-video, exactly one image selects image-to-video, and one to three videos select video editing; images and videos cannot be mixed. The default is 1080p, 5 seconds, shot structure Off, and content filtering On.

Review Wan 2.5 Video on Kyeo AI with text-to-video or one-image input, 5/10 seconds, 720p/1080p, prompt constraints, and 60–200-credit pricing. Kyeo charges by resolution and duration: 60 credits for 720p at 5 seconds, 120 for 720p at 10 seconds, 100 for 1080p at 5 seconds, and 200 for 1080p at 10 seconds. The 720p 5-second default costs 60 credits. How are Wan 2.5's 60/100/120/200 credits calculated? — Review Wan 2.5 Video on Kyeo AI with text-to-video or one-image input, 5/10 seconds, 720p/1080p, prompt constraints, and 60–200-credit pricing.

Wan 2.2 A14B Turbo is not a single Kyeo workflow. No media selects text-to-video, one image selects image-to-video, and one image plus one audio file selects speech-to-video. Each workflow has different resolutions, credit rates, and visible controls, so confirm the media combination first.

Wan 2.2 Animate Replace uses one source video and one replacement image uploaded through the workbench, with 480p, 580p, or 720p output. Pricing uses verified source-video seconds, and users cannot enter arbitrary media URLs directly.

Kyeo has six valid combinations: Standard 768P at 6/10 seconds and 1080P at 6 seconds cost 25/40/40 credits; the matching Pro prices are 40/80/70. 1080P does not support 10 seconds. Exactly 1 JPG/PNG/WebP file, up to 10MB.

Hailuo Pro / Workflows — This page is based on official MiniMax materials for Hailuo 02 and the product contract currently verified by the site. It covers only the available text and image workflows, verifiable limits, and credit prices. Kyeo has one Pro combination: text and image requests both target 1080P at six seconds and cost 57 credits; adding an end frame does not add a separate charge.

Hailuo Standard combines text-to-video with first-frame animation and an optional last frame. Five priced combinations range from 12 to 50 credits.

Review Kling 3.0 on Kyeo AI: 3–15 seconds, single-shot or up to 5 multi-shots, first/last frames, sound, and std/pro/4k per-second rates. A fit for specifying first and last frames or splitting a 3–15-second concept into up to 5 short shots. Character consistency, lip sync, text, sound, and shot continuity still require review of the actual result.

What inputs does Kling 3.0 Motion Control require? — 1 image + 1 video — Billing uses the signed motion-video duration: 20 credits per second for `720p` or 27 per second for `1080p`, multiplied by the verified 3–30-second duration and rounded up. Missing or invalid duration rejects generation.

Kling 2.6 is a two-workflow video service on Kyeo: text requests use text-to-video, while requests with an image use single-image animation. Both routes offer 5/10 seconds and a `sound` toggle; text generation also provides three aspect ratios. The four Kyeo credit tiers are 55, 110, 110, and 220.

Kling 2.6 Motion Control: one image and one motion video, orientation limits, 3–30 seconds, and per-second rates of 11 credits at 720p or 18 at 1080p. It can test whether single-character motion transfer fits your image and video. Body, face, hands, identity, background, text, and fast-motion behavior still require trials with your own assets.

Kling V2.5 Turbo Pro gives Kyeo separate text and image routes behind one entry. Uploading 1-2 images switches to first/last-frame video; both routes offer 5 or 10 seconds at 42 or 84 credits.

Kling V2.1 Master is Kyeo's 1080p text and single-image video entry. No image selects the text workflow; 1 JPEG/PNG selects the image workflow. Both cost 160 credits for 5 seconds or 320 for 10.

Kling V2.1 Pro is Kyeo's High Quality (1080p) first-last-frame image-to-video route. It requires 1 first image and can use image 2 as the ending; 5- and 10-second requests cost 50 and 100 credits.

Kling V2.1 Standard is Kyeo's 720p single-image video route. It requires 1 JPEG or PNG plus a prompt and returns a 5- or 10-second task for 25 or 50 credits, respectively.

Workflow: The UI requires exactly 1 JPEG or PNG image and 1 MPEG, WAV, X-WAV, AAC, MP4, or OGG audio file. The image may be up to 10 MB; audio may be up to 100 MB and 5 minutes. The charge uses the signed audio duration at 16 credits per second and rounds up. A missing, invalid, or over-5-minute duration cannot be submitted.

Workflow: The UI requires exactly 1 JPEG or PNG image and 1 MPEG, WAV, X-WAV, AAC, MP4, or OGG audio file. The image may be up to 10 MB; audio may be up to 100 MB and 5 minutes. The charge uses the signed audio duration at 8 credits per second and rounds up. A missing, invalid, or over-5-minute duration cannot be submitted.

Infinitalk uses 1 image, 1 audio file no longer than 15 seconds, and a required prompt to generate a speaking-portrait video on Kyeo. It offers 480p or 720p plus optional `seed`; cost depends on audio duration and resolution rather than a fixed request price.

Runway currently supports text-to-video and single-image-to-video. Base generation offers 720p for 5 seconds, 720p for 10 seconds, and 1080p for 5 seconds.

Runway — Does Runway Aleph support a reference image? — 90 credits / request. Use it to request additions, removals, replacements, relighting, restyling, or content changes to existing footage. Use it to request additions, removals, replacements, relighting, restyling, or content changes to existing footage.

Does Kyeo AI currently use Grok Imagine Video 1.5 Preview? Platform route selection — 1–15 seconds; 480p or 720p; up to 7 reference images

Topaz Video Upscaler is Kyeo's current single-video factor workflow. It accepts one video URL plus `1x / 2x / 4x`, defaults to `2x`, and has no prompt input. Those factors are API values, not guarantees of denoising, restoration, stabilization, frame interpolation, 4K output, or a specific visual improvement; test critical footage first.

Veo 3.1 Quality on Kyeo accepts text or one to two first/last-frame images through the public source Veo 3.1 Quality route for 225 credits. Eligible results can launch Extend, 1080p, or 4K actions separately. Quality is a workflow name, not a fixed delivery guarantee.

Veo3.1 Fast on Kyeo AI supports text, first/last-frame and reference-image generation, direct 720p/1080p/4K selection, and 4/6/8-second duration controls. 30 / 38 / 90 credits by resolution.

Veo3.1 Lite on Kyeo AI supports text, first/last-frame and reference-image generation, direct 720p/1080p/4K selection, and 4/6/8-second duration controls. 15 / 23 / 75 credits by resolution.

Kling 3.0 Turbo targets 3–15-second video generation. It uses text mode when no image is uploaded and switches to image-to-video when one valid image is supplied. Resolution and duration directly determine the preauthorization amount.

Seedance 2.0 Mini combines text, first/last-frame, and multimodal-reference generation in one entry point. Reference video enters the pre-task pricing formula, so both media duration and output duration must be verified before task creation.

Seedance 2.5 combines a 30-second boundary, higher media counts, and video-editing reference input in one strict scene contract. Automatic duration is not underestimated at the 5-second default; Kyeo preauthorizes against the 30-second maximum.

MiniMax H3 offers 768P and 2K output with text, first/last-frame, and image/video/audio reference scenes. Reference-video duration and every input image from image 6 onward affect pre-task authorization.

HappyHorse 1.1 routes by image count: no image selects text mode, 1 image selects first-frame image-to-video, and 2–9 images select reference mode. All three modes share the same resolutions, durations, and per-second rates.

The current OmniHuman 1.5 core path turns one image of a person, pet, or animated subject plus one audio clip into a driven video. Signed upload metadata verifies audio duration, which directly determines pre-task authorization.

Video Lip Sync rebuilds mouth movement in one existing video to match a target vocal track. Lite mode targets front-facing single-person video, while Basic mode handles more complex single-person scenes and can enable scene detection.

Wan 3.0 combines Standard and Prime tiers with text, first/last-frame, and signed multimodal-reference generation. The reservation is the variant-and-resolution rate multiplied by verified input-video seconds plus reserved output seconds, rounded up once. Intelligent duration reserves 30 output seconds and is unavailable with reference video.

Current status — Verified single-shot generation is available from text, one first frame, or a first-and-last-frame pair. Video reference, transformation, multi-shot, and element workflows remain closed because those contract branches are still ambiguous.