Unverified50% confidenceBenchmarkExact time
In testing, a video with 1344P resolution and 15 seconds duration was generated, containing both visuals and audio.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/3/2026
First Seen
Valid until: 1/1/2027
Sources
Related Entities
Related Claims
UnverifiedIn video generation, the computational cost of self-attention is proportional to the square of the number of tokens; after VAE encoding and patching, a 4-second 720p 24fps video can yield tens of thousands to over a hundred thousand tokens in total.77% similarUnverifiedThe model contains approximately 20.8 million parameters, with channel counts progressing as 96→192→384, and supports generating 16-frame video clips at a resolution of 64×64.76% similarUnverifiedAtlas can generate video content with a resolution of up to 1440p and a duration of up to one minute75% similarVerifiedThis workflow can generate continuous videos up to 60 seconds long on a 12 GB GPU.74% similarUnverified16GB of video memory sits in a critically usable range among current mainstream video diffusion models, and some high-resolution or long-sequence models still experience out-of-memory issues.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/970312API
curl https://kongchang.com/api/v1/knowledge/claims/970312MCP
get_claim(id=970312)