ShengShu Technology releases Vidu Q4 Preview, ranking 3
BIBIBI
AT A GLANCE
ShengShu Technology releases Vidu Q4 Preview, ranking 3。
Article
ShengShu Technology's Vidu Q4 breaks into the top for image-to-video: its previous generation ranked 19 Beating News AI reports that ShengShu Technology has released its new-generation flagship video model, Vidu Q4 Preview, now available on the web and via API. The new model currently supports image-to-video and reference-to-video generation, but not text-to-video generation.
On Artificial Analysis's blind evaluation leaderboard for image-to-video (including audio), the new model ranked 1179 with a score of 3, behind only MiniMax H3 Max and MiniMax H3. The previous-generation Vidu Q3 Pro scored 1056 points and ranked 19. The generations differ by 123 points.
Vidu Q4 Preview primarily improves facial expressions, body movements, camera transitions, and visual effects such as explosions. It can generate visuals and sound simultaneously, with a maximum single-video length of 16 seconds and support for up to 4K resolution. Users can also upload up to 15 reference images and 3 audio references to specify character appearance, scenes, and voices, helping characters remain as consistent as possible across different shots.
Original link https://m.theblockbeats.info/flash/370806
Key points
01
ShengShu Technology has released its new-generation flagship video model, Vidu Q4 Preview, now available on the web and via API.
02
Vidu Q4 Preview supports image-to-video and reference-to-video generation, but not text-to-video generation.
03
On Artificial Analysis's blind evaluation leaderboard for image-to-video (including audio), Vidu Q4 Preview ranked 1179 with a score of 3。
04
MiniMax H3 Max and MiniMax H3 rank ahead of Vidu Q4 Preview.
05
The previous-generation Vidu Q3 Pro scored 1056 points and ranked 19, with a difference of 123 points between the generations.
06
Vidu Q4 Preview improves facial expressions, body movements, camera transitions, and visual effects such as explosions.
07
The model can generate visuals and sound simultaneously, with a maximum single-video length of 16 seconds and support for up to 4K resolution.
08
Users can upload up to 15 reference images and 3 audio references to specify character appearance, scenes, and voices.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
Vidu Q4 Preview is a video generation model released by Shengshu Technology. It currently mainly generates videos from images and reference videos, and can also generate sound simultaneously, but cannot yet generate videos from text alone. Reported blind-test results show that it scored higher and ranked above the previous-generation model.
Why it matters to readers
This indicates that Vidu Q4 Preview ranked higher than Vidu Q3 Pro in the blind evaluation described in the report and offers capabilities including video duration, resolution, audio, and multiple reference inputs.
If you want to create videos from images, pay attention to the supported input types, maximum duration, resolution, and number of reference assets.
Risks and unknowns
The blind-test ranking and score were relayed in reports from the Artificial Analysis leaderboard; the original source did not provide the leaderboard's original test materials.
The original source did not specify the detailed pricing, usage limits, and availability scope for the web version and API.
The original source did not specify whether 4K、16 seconds, multiple reference images, and sound references are available through all usage channels.
Related Developments
Loading event timeline…
Related concepts
Image-to-video
This term is not in the glossary yet. Browse related concepts in the glossary.
