AIVeed added four things on October 29, 2026: a cheaper Veo 3.1 Lite tier with 1080p, 4K and 4-, 6- or 8-second clips; Kling 3.0 on the Omni engine with up to 6 shots per clip; Seedream 5 images and a Remove background tool; and Avatar & Dub, which makes talking-avatar videos of up to 5 minutes and dubs videos into any language (AIVeed pricing, October 29, 2026).
What changed in Veo 3.1?
Veo 3.1 now has three tiers: Lite, Fast and Quality (in the German interface, Lite, Schnell and Qualität). You choose 720p, 1080p or 4K, and a length of 4, 6 or 8 seconds; the length does not change the price, only the tier and the resolution do. Lite and Fast also accept up to 3 reference images (8-second clips) as an alternative to a first and last frame. Veo clips still come with sound, and an existing clip can still be extended at the price of the tier it was made with: Lite 60, Fast 120, Quality 500 credits. Quality used to cost 320 credits and now costs 500 at 720p.
| Tier | 720p | 1080p | 4K |
|---|---|---|---|
| Lite | 60 ($0.30) | 70 ($0.35) | 300 ($1.49) |
| Fast | 120 ($0.59) | 130 ($0.64) | 360 ($1.78) |
| Quality | 500 ($2.48) | 510 ($2.52) | 740 ($3.66) |
The free 200 credits that come with every new account now buy three Veo 3.1 Lite videos, up from one Fast video, which makes Lite the cheapest way to try Veo. For Google's own per-second rates next to these credits, see our Veo 3 pricing guide; the Veo3.1 model page has the controls.
What is new in Kling 3.0?
The Kling 3.0 tab now runs on Kling 3.0 Omni at the same prices. Multi-shot prompts now allow up to 6 shots in one clip (it was 5), each 1 to 12 seconds, 15 seconds in total. Aspect ratios are 16:9, 9:16 and 1:1, and character or subject references work as before.
| Resolution | 6 s | 10 s | 15 s |
|---|---|---|---|
| 720p | 225 | 300 | 450 |
| 1080p | 300 | 450 | 600 |
What are Seedream 5 and Remove background?
Seedream 5 is ByteDance's image model, from the same lab as Seedance, which makes it a natural start frame for a Seedance video. It generates and edits images in two versions: Seedream 5 Pro at 15 credits (1K) or 30 credits (2K), and Seedream 5 Flash at 8 credits (1K or 2K). When the video model is Seedance, the "Generate first frame" image tool opens on Seedream 5. In My Images, Remove background makes a transparent PNG copy of any image for 3 credits, which is handy for product shots.
What is Avatar & Dub?
Avatar & Dub is a separate page (not a tab in the generator) for two jobs. A talking avatar turns one photo and a voice into a presenter video that talks and gestures, from 4 seconds to 5 minutes at 720p, for 28 credits per second of voice. Dub takes any video of someone speaking plus a new voice and returns the same video speaking the new audio, up to 4 minutes, for 16 credits per second, in a "one speaker facing the camera" or a "cuts or several speakers" mode. The lips follow the audio, so it works in any language the voice speaks.
| Length | Talking avatar (28 per second) | Dub (16 per second) |
|---|---|---|
| 30 seconds | 840 | 480 |
| 60 seconds | 1,680 | 960 |
For the voice you either write a script and pick a voice, which adds voiceover credits (20 for up to 15 seconds, 40 up to 30 seconds, 80 up to 60 seconds, 150 up to 2 minutes, 300 up to 5 minutes), or upload your own MP3 or WAV. The existing Lip Sync tab stays (480p from 40 credits), and links to the new page; every finished video in My Videos also has a "Dub into another language" button.
How do you make a talking video in German?
Take a photo of your presenter, write the script in German, and pick a German voice. A 30-second German presenter clip then costs 840 credits for the avatar plus 40 for the voiceover, 880 credits in total ($4.36 on the Business pack); on the German interface the page is called "Avatar & Vertonung" (aiveed.io/de/avatar). To localise a video you already have, dub it instead: a 60-second video re-voiced into German costs 960 credits, plus 80 if the voice comes from a written script. The same route works for French, Spanish, Italian, Dutch, Polish and other languages. We recommend Avatar & Dub for EU-language talking videos rather than relying on the native audio of a video model such as Veo, Kling or Seedance, because we have not tested lip accuracy in German for those.
For the background to this problem, see why most AI talking-head videos fail.
Is the API updated too?
Yes. The model ids are veo3-lite, veo3-fast and veo3-quality (with resolution of 720p, 1080p or 4k, duration of 4, 6 or 8, and up to 3 reference_image_urls on Lite and Fast), kling-3.0 (Omni engine, up to 6 entries in multi_prompt), seedream-5-pro and seedream-5-flash (1K or 2K), avatar (image_url and audio_url) and dub (video_url and audio_url, mode lite or basic) (AIVeed API reference, October 29, 2026).
Try a Veo 3.1 Lite video from 60 credits, or open Avatar & Dub, on AIVeed. New accounts start with 200 free credits.