# NVIDIA Cosmos 3 Super serving tradeoffs on Nebius, latency vs throughput

> Original title: Serving World Models: Latency vs Throughput with NVIDIA Cosmos 3 Super on Nebius | Cosmos Labs https://x.com/i/broadcasts/1mxPaZvloNLKN

- Source: NVIDIA AI
- Published: 2026-10-28T16:02:17.000Z
- Bad Influence: https://badinfluence.news/items/s39we8qtke81i5456exxpjzv6
- Original: https://x.com/NVIDIAAI/status/2115474241894416418

## Summary

NVIDIA's Cosmos Labs broadcast covers serving world models with Cosmos 3 Super on Nebius, comparing latency and throughput. The post links to a live broadcast and provides no performance figures.

## Post

Serving World Models: Latency vs Throughput with NVIDIA Cosmos 3 Super on Nebius | Cosmos Labs https://x.com/i/broadcasts/1mxPaZvloNLKN