Models
Video Models
Every video generation and editing model available on Kolbo.AI
Kolbo.AI offers the full range of modern video models for text to video, image to video, reference-driven generation, native audio, and natural-language video editing.
Available Models
| Model | Provider | Modes | Resolution / Duration | Best for |
|---|---|---|---|---|
| LTX 2.3 Quality | Lightricks | Text to video, image to video, Ingredients (up to 9 refs), native audio | Up to 4K, 5 to 20s | Cinematic short films and branded content combining many reference elements in one shot |
| Luma Ray 3.2 | Luma | Text to video, image to video, first-last frame, reframe, video to video | Up to 1080p, 5 or 10s, square to 21:9 | Film-grade cinematic B-roll and social clips with flexible framing |
| HappyHorse 1.1 | Alibaba | Text to video, image to video, reference to video (up to 9 refs), native audio | 720p / 1080p, 3 to 15s | Character-consistent video across clips at a practical 1080p cost |
| HappyHorse 1.0 | Alibaba | Text to video, image to video, video edit (reference to video) | 720p / 1080p, 3 to 15s | Earlier-generation character and style consistent video editing |
| Kling 3.0 | Kuaishou | Text to video, image to video, multi-shot | Up to 4K, up to 15s multi-shot | Native-4K multi-shot cinematic sequences with character consistency |
| Kling O1 | Kuaishou | Video to video VFX, character swap, style transfer | 1080p / 30fps, up to 2 minutes | One-click AI VFX that reimagines footage while preserving original motion |
| Kling O3 | Kuaishou | Image generation, natural-language video edit (Standard and Pro) | Up to 4K, 3 to 15s | Character-consistent 4K storytelling and conversational video editing |
| Seedance 2 | ByteDance | Text to video, image to video, first-last frame, Elements | Multiple aspect ratios, 4 to 15s | Sharp, prompt-accurate cinematic video across all video tools |
| Seedance 1.5 | ByteDance | Text to video, image to video, native audio | Synchronized sound built in | First Seedance with built-in synchronized sound out of the box |
| Gemini Omni Video | Text to video, image to video, multimodal input, native audio | 720p / 1080p / 4K, 4 to 10s | Multimodal narrative and character-driven shots with voice and character consistency | |
| Gemini Omni Flash | Multimodal input (text, image, video, voice), conversational refinement | Multimodal | Any-input multimodal video with step-by-step conversational editing | |
| Google Veo 3 | Text to video, image to video, first-last frame, reference to video, native audio | Up to 1080p | Cost-efficient Google video generation with native audio | |
| Hailuo 2.3 | Minimax | Text to video, image to video | Cinematic short clips | Smooth, expressive motion and reliable prompt following |
| Sora 2 | OpenAI | Text to video, image to video (Standard and Pro) | Up to 1080p, 4 to 20s | Premium cinematic generation with strong physical realism |
Notes
- LTX 2.3 Quality Ingredients mode lets you combine up to 9 reference assets in a single shot, with native audio always included.
- HappyHorse 1.1 adds native audio and sharper 1080p output over the 1.0 generation, at a lower cost.
- Kling spans three models: 3.0 for native-4K multi-shot generation, O1 for one-click video VFX up to two minutes, and O3 for image generation plus natural-language video editing.
- Seedance 1.5 is the first Seedance with built-in synchronized audio. Seedance 2 is the sharper, more prompt-accurate flagship.
- Gemini Omni comes in two forms: Video (seeded with native audio, up to 4K, 4 to 10s) and Flash (multimodal any-input with conversational editing).
Where to use these
These models power the Video Presets, Image to Video, Text to Video, Video to Video, Lipsync, Elements, First Last Frame, and the cinematic preset tools. See Video Tools for the full set of workflows.