Picture and sound in one pass
Kling AI presents version 2.6 as its first model that generates picture and sound at the same time. Turn on “Generated sound” for voices, sound effects and ambience synced with the action.
Video model · Kling AI
Kling 2.6 — 5 or 10 s clips, with sound when you want it.
Kling 2.6 (Kling VIDEO 2.6) is the video model from Kling AI (Kuaishou), launched in December 2025 as the brand’s first model to generate picture and sound together. On Deepnia it makes 5 or 10 second clips from text or a photo; turn on “Generated sound” to add voices, sound effects and ambience, with characters who speak or sing in English or Chinese.
- 5 or 10 s
- Generated sound
- Text or 1 image
- 9:16 · 1:1 · 16:9
- Voices, singing, sound effects
The video tool opens with Kling 2.6 already selected. The exact cost shows before every creation; you sign up when you generate.
- Your prompt
- A start image (optional)
- Generated sound: on or off
- A 5 or 10 s clip
At a glance
- Maker
- Kling AI (Kuaishou)
- Launched
- December 2025
- Length
- 5 or 10 s (5 s by default)
- Aspect ratios
- 9:16, 1:1, 16:9
- Start image
- 1, optional
- Sound
- Voices, singing, sound effects and ambience (“Generated sound”)
What Kling 2.6 does best
Voices, dialogue and even singing
Kling documents speech, dialogue, narration, singing and rap, on top of sound effects and ambience. Characters speak and sing in English or Chinese, so write your lines in one of those two languages.
Animate your own photo
Add 1 image: it becomes the video’s first frame, and the clip keeps its aspect ratio. A dish, a shop window or a product comes to life, with or without sound.
Simple settings
A choice of 5 or 10 seconds, 3 social formats (9:16, 1:1, 16:9) and one switch for sound: Kling 2.6 gets straight to the point when a short clip is enough.
How to use Kling 2.6 on Deepnia
Open the video tool
The “Try Kling 2.6” button opens Deepnia’s video tool with the model already selected.
Add a photo if you want
Add 1 image of up to 10 MB to use as the start frame. Without an image, the video starts from your text.
Write the prompt
Describe the scene, the action and the camera move, then the sound you want: ambience, effects and lines of dialogue in quotation marks.
Set it up and generate
Pick 5 or 10 seconds and the aspect ratio (without an image), turn on “Generated sound” to add sound, check the cost shown, then create.
Kling 2.6 on Deepnia: what you can set
- Inputs
- Text only, or text and 1 start image
- Length
- 5 or 10 s (5 s by default)
- Aspect ratios
- 9:16, 1:1, 16:9 without an image; with an image, the video keeps its aspect ratio
- Start image
- 1 at most, up to 10 MB
- Sound
- With “Generated sound”: voices, singing, sound effects and ambience
- Voice languages
- English and Chinese
- Prompt
- Required; stay under 2,500 characters, Kling’s documented limit
Prompt ideas to try
Copy a prompt, open the tool and adapt it to your project.
Animated dish for a restaurant
- 1 image: photo of the dish
- 5 s
- Sound on
Animate this photo: steam rises gently from the dish, the sauce glistens, a hand places a spoon next to the plate, slight push-in. Sound: soft sizzling, restaurant ambience, clinking cutlery.
Sung jingle for a juice brand
- Text only
- 10 s
- 1:1
- Sound on
A smiling young woman holds a bottle of mango juice in front of a colorful fruit stall and sings a short cheerful line: "Fresh mango, sunny day!" Sunny light, gentle circling camera. Sound: upbeat pop singing, market ambience.
Pep talk from a fitness coach
- Text only
- 10 s
- 9:16
- Sound on
In a bright gym, a coach in sportswear faces the camera while her group does squats behind her. [Coach, energetic]: "Ten more seconds, you've got this!" Medium shot, handheld camera. Sound: breathing, footsteps, gym ambience.
Shop window for a Lagos fashion boutique
- 1 image: photo of the window
- 10 s
- Sound on
From this photo of the shop window, the camera slowly moves toward the mannequins, the fabrics sway slightly, the window lights switch on one by one. Sound: distant street noise, passing footsteps, a soft door chime.
Silent product shot for an online store
- 1 image: photo of the watch
- 5 s
- Sound off
The camera slowly circles the watch on a stone pedestal, soft reflections on the dial, dark background, studio lighting, smooth and steady motion.
Animated holiday card for a company
- Text only
- 5 s
- 16:9
- Sound on
An office decorated for the holidays: a string of lights switches on, gold confetti falls on a table covered with wrapped gifts, slow sideways tracking shot. Sound: sleigh bells, distant laughter, rustling gift wrap.
Tips for better results
- Describe the sound at the end of the prompt (ambience, effects, voice or singing) so it matches the action.
- For dialogue, give each character a unique label and write each line as [Character, emotion]: "line", the format Kling recommends.
- Write lines in English or Chinese, the two languages of Kling 2.6’s voices.
- Choose 10 seconds for dialogue or a song: that is the length Kling recommends.
- Start from a sharp, well-defined photo: Kling notes that image-to-video quality depends heavily on the input image’s resolution.
- With a start image, describe the motion and the camera rather than what the photo already shows.
Good to know
- Pick 9:16 for Reels, TikTok and Shorts, 16:9 for YouTube and 1:1 for a square feed; with a start photo, crop it to the right format before adding it.
- For characters who speak French, choose Gemini Omni; for a French voice-over, create it with Open TTS, Deepnia’s voice model, then add it in your editor.
- Start from photos you have the rights to, especially to animate a real person.
Kling 2.6 or another model?
Deepnia’s other video models, to pick the right one for your project.
| Model | Length | Resolution | Aspect ratios | Best for |
|---|---|---|---|---|
| Kling 2.6This model | 5 or 10 s | Automatic | 9:16 · 1:1 · 16:9 | 5 or 10 s clips, with sound when you want it. |
| Kling 3.0 | 3–15 s | Set by the mode | 9:16 · 1:1 · 16:9 | 3–15 s, up to 5 shots, 3 persistent elements and an end frame. |
| Motion Control | Set by the source video | 720p–1080p | From the source | Copying the moves from a real video onto a character image. |
| Seedance 2.0 | 4–15 s | 480p–1080p | 16:9 · 9:16 · 1:1 · 3:4 · 4:3 · 21:9 | Videos with sound up to 1080p, guided by image and video references. |
| Veo | 8 s | Set by the mode | 9:16 · 16:9 | 8 s clips with sound, in 3 quality levels. |
| HappyHorse | 3–15 s | 720p–1080p | 9:16 · 1:1 · 16:9 · 4:3 · 3:4 | Editing an existing video or combining up to 9 images. |
Frequently asked questions
How long is a Kling 2.6 video?
5 or 10 seconds, 5 seconds by default. Kling recommends 10 seconds for dialogue or a song.
Does Kling 2.6 generate sound?
Yes: turn on “Generated sound” and Kling 2.6 creates voices, dialogue, singing, sound effects and ambience together with the picture. Describe the sound you want at the end of the prompt.
How do I make a character speak or sing?
Turn on “Generated sound”, then write the line in English or Chinese as [Character, emotion]: "line", the format Kling recommends. Choose 10 seconds for dialogue or a song.
Can I animate my own photo?
Yes: add 1 image (up to 10 MB). It becomes the start frame and the video keeps its aspect ratio.
Which aspect ratios are available?
Without an image: 9:16, 1:1 or 16:9. With an image: the photo’s aspect ratio.
How do I get a longer video?
Create several 10-second clips and join them in your editor. To chain up to 5 shots in one 15-second video, use Kling 3.0.
How do I keep the same product from clip to clip?
Start from the same product photo for each creation and describe it with the same words. To keep up to 3 characters or products identical within one video, use Kling 3.0’s persistent elements.
Kling 2.6 or Kling 3.0: which one should I use?
Kling 2.6 for a 5 or 10 second clip that is quick to set up, from text or a photo. Kling 3.0, launched in February 2026, for any length from 3 to 15 seconds, up to 5 shots, 3 persistent elements and start and end frames.
Ready to try Kling 2.6?
The video tool opens with Kling 2.6 already selected. The exact cost shows before every creation; you sign up when you generate.