UC Berkeley's VDN-H3 Generates 14-Second Videos Faster Than You Can Watch

VDN-H3 replaces most of MiniMax H3's quadratic attention with a linear branch, rendering a 14.4-second 768p clip in 11.23 seconds on 8 B200s.

Source

AlphaSignal

Published

Sep 7, 2026, 7:56 AM

Author

AlphaSignal Newsroom

Read

1 min read

VDN-H3 replaces most of MiniMax H3's quadratic attention with a linear branch, rendering a 14.4-second 768p clip in 11.23 seconds on 8 B200s.

Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.

Next reads