Kyutai released CADRE, a family of 7 open causal diffusion video autoencoders for interactive world models. World models built on CADRE run at 20 fps on 1 H100 and react in under 230 ms
Kyutai CADRE: Video Autoencoders That Cut World Model Lag Below 230 ms
Kyutai released CADRE, a family of 7 open causal diffusion video autoencoders for interactive world models. World models built on CADRE run at 20 fps on 1 H100 and react in under 230 ms
M
Michal Sutter · via MarkTechPost
Curated & Verified Aggregation
Friday, October 23, 2026
2 min
Original reporting by MarkTechPost
Read the full story at MarkTechPostListen to this story
Narrated edition • 2 min read time
researchaitalky dispatch
“Kyutai CADRE: Video Autoencoders That Cut World Model Lag Below 230 ms”
via MarkTechPostPRIMARY SOURCE
Related Stories
Forced to pick, Grok's AI model chooses Republicans. OpenAI's picks Democrats
The PHP Meltdown Shows How AI Makes Bugs More Exploitable
researchaitalky dispatch
“Meet Anthropic's sandbox-runtime: OS-Level Sandboxing for AI Agents, No Containers”
via MarkTechPostPRIMARY SOURCE