Action motion
Repeat a complex movement prompt three times and count failures.
Gemini Omni Flash combines multimodal references with conversational generation and editing. Kuaishou's Kling 3.0 family adds text, image, audio and video input, native multilingual audio, reference-to-video, in-video editing and multi-shot storyboards up to 15 seconds.
Official Google and Kuaishou documentation checked September 8, 2026

Publicly documented capabilities are compared separately from subjective output quality.
Last updated
Sources: Gemini Omni guide and pricing; Kuaishou's February 5, 2026 Kling 3.0 launch release. Kuaishou's release does not publish an API list price or video resolution table, so those rows describe disclosure rather than assume a value.
| Feature | Gemini Omni Flash | Kling 3 |
|---|---|---|
| Official access | Google AI Studio / Gemini API | Kling AI product; access described in Kuaishou's official launch |
| Input / reference modes | Text, image and short-video references; first/last frame | Text, image, audio and video; multiple images and reference video |
| Duration | 3–10 seconds; extension up to 40 seconds total | Up to 15 seconds |
| Native audio | Video with generated audio | Multilingual voices, dialects and accents; sound effects and music |
| Editing / storyboard | Conversational edits and scene extension | In-video editing and multi-shot storyboard generation |
| Resolution disclosure | 360p, 720p, 1080p or 4K | No video resolution table in the official launch release |
| Official price disclosure | $0.10 per output second | No API list price in the official launch release |
Documented capability coverage
A transparent inference from official disclosures, not a claim about subjective output quality.
Scoring: 10 = broad, explicit official support; 5 = partial or constrained disclosure; 0 = absent from the cited official material. Resolution and pricing scores measure public evidence coverage, not hidden product capability.
Evidence snapshot: official documentation checked 2026-09-08
● Test protocol
These illustrations describe the test scenes; they are decorative and are not model outputs.
Repeat a complex movement prompt three times and count failures.
Measure subject and composition drift from the supplied first frame.
Kling 3.0's official launch documents broader native input types, multi-shot storyboards and clips up to 15 seconds. Gemini Omni Flash documents 4K, transparent list pricing, conversational editing and extension up to 40 seconds total. The chart compares documented coverage, not visual quality.
● FAQ
Generate with pay-as-you-go credits and no monthly commitment.