Kling 2.6

Kling 2.6, officially Kling VIDEO 2.6, is Kuaishou's AI video model released in December 2025. It was the first Kling model to generate audio and video together — voice-over, sound effects and ambience in one pass — and makes 5 or 10 second clips from a text prompt or a start image. On API platforms it appears as kwaivgi/kling-v2.6-pro and kwaivgi/kling-v2.6-std.

On Kenerate, Kling 2.6 Pro has the sound toggle and takes a start and an end frame, for 21 credits a second silent or 42 with sound. Kling 2.6 Standard is silent, takes a start frame only and costs 12.6 credits a second. Both make 5 or 10 second clips in 16:9, 9:16 or 1:1 from text or an image.

It's billed per second. A 5-second clip costs 63 credits on Standard, 105 on Pro silent and 210 on Pro with sound; a 10-second clip costs 126, 210 and 420. Credits come in one-time packs that never expire, starting at $15 for 1,300 credits, with no subscription.

Yes — native audio is the headline feature. Kling's guide says VIDEO 2.6 generates voice-over, sound effects and ambient sound that line up with the action, with voice output in Chinese and English (other languages are translated to English). On Kenerate, sound is a toggle on Kling 2.6 Pro; Standard renders silent.

English and Chinese. Kling's 2.6 guide states the model currently supports Chinese and English voice output and auto-translates other languages to English. If you need dialogue in Japanese, Korean or Spanish, Kling 3.0 adds those.

5 or 10 seconds per clip — nothing in between. Kling recommends 10 seconds for dialogue and singing so lines aren't rushed. For longer pieces, generate several clips, start each from the last frame of the one before, and join them in an editor.

Not on Kenerate's free plan: video generation needs a paid plan, which means buying any one-time credit pack. New accounts get 25 starter credits for images. Kling's own app has its own free allowance and paid tiers; check kling.ai for its current terms.

Yes. Upload a start image in /app/video and Kling 2.6 animates it — a short motion prompt is required. Kling 2.6 Pro also accepts an end image and moves from the first frame to the last, but with an end frame the clip comes out silent. Image to video is the surest way to keep a product or character exactly as designed.

It's the API name for Kling 2.6 Pro on model platforms such as WaveSpeed — kwaivgi is Kuaishou's Kling namespace, -std is the Standard tier, and there are separate text-to-video and image-to-video endpoints. On Kenerate you don't need an API key: the same model runs in the browser as Kling v2.6 Pro.

Kling 3.0 makes 3–15 second clips, adds dialogue in Japanese, Korean and Spanish and has a 4K tier, but costs more: 25.2–33.6 credits a second silent on Kenerate. Kling 2.6 is the cheaper choice for 5 or 10 second clips in English or Chinese — 12.6 a second on Standard, 21 on Pro silent.