Kling Models by Kuaishou Technology
Last refreshed 2026-11-04. Next refresh: weekly.
Details
Capabilities
About
Kling (可灵) is Kuaishou Technology's family of AI video and image generation models, offered through the Kling AI creative studio and the KlingAI Open Platform API. Kuaishou built the first Kling video model on a diffusion transformer (DiT) with its own 3D VAE, turning text prompts into clips of up to two minutes at up to 1080p and 30fps. Later updates added image-to-video, video extension, start-and-end-frame control, and shot control. The Kling 3.0 series (Video 3.0, Video 3.0 Omni, Image 3.0, and Image 3.0 Omni) supports clips of up to 15 seconds, native audio in multiple languages, and text-to-video, image-to-video, reference-to-video, and in-video editing in one multimodal model.
Decision facts
- Best fit
- video generationvision and multimodal workagent workflows
- Capability starting point
- Kling 4.0 Flash with multimodal inputs
- Lowest tracked input
- Not tracked
- Closest related family
- Kolors
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs video generation, multimodal inputs, and audio.
Use when the workload needs video generation, multimodal inputs, and audio.
Use when provider availability and model metadata match the workload.
Use when provider availability and model metadata match the workload.
Use when provider availability and model metadata match the workload.
Use when provider availability and model metadata match the workload.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Kling 4.0 Flash | Use when the workload needs video generation, multimodal inputs, and audio. | 2026-11 | video generationmultimodal inputsaudio | Current |
| Kling 4.0 | Use when the workload needs video generation, multimodal inputs, and audio. | 2026-10 | video generationmultimodal inputsaudio | Current |
| Kling IMAGE 3.0 | Use when provider availability and model metadata match the workload. | 2026-02 | — | Current |
| Kling VIDEO 3.0 Pro | Use when provider availability and model metadata match the workload. | 2026-02 | — | Current |
| Kling VIDEO 3.0 Standard | Use when provider availability and model metadata match the workload. | 2026-02 | — | Current |
| Kling IMAGE O3 | Use when provider availability and model metadata match the workload. | 2025-02 | — | Current |
| Kling VIDEO O3 Pro | Use when provider availability and model metadata match the workload. | 2025-02 | — | Current |
| Kling VIDEO O3 Standard | Use when provider availability and model metadata match the workload. | 2025-02 | — | Current |
| Kling | Use when the workload needs multimodal inputs. | 2024-06 | multimodal inputs | Current |
Release Timeline
5 release groupsSpecifications(9 models)
| Model | Released | Multimodal |
|---|---|---|
| Kling 4.0 Flash | 2026-11 | Yes |
| Kling 4.0 | 2026-10 | Yes |
| Kling IMAGE 3.0 | 2026-02 | No |
| Kling VIDEO 3.0 Pro | 2026-02 | No |
| Kling VIDEO 3.0 Standard | 2026-02 | No |
| Kling IMAGE O3 | 2025-02 | No |
| Kling VIDEO O3 Pro | 2025-02 | No |
| Kling VIDEO O3 Standard | 2025-02 | No |
| Kling | 2024-06 | Yes |
