From the Lab: Text Diffusion and Elastic Reasoning | Nemotron Labs

Nemotron Labs is where NVIDIA researchers work on what open models can do next and in this livestream, we hear directly from the people building them. We'll share the latest on two research-stage models: Nemotron Labs Diffusion, a text-native diffusion model exploring a fundamentally different approach to language generation, and Nemotron Labs Elastic, a many-in-one reasoning model that packs three nested model sizes into a single checkpoint with dynamic compute budget control. The researchers will give updates on both, including what's still in progress, and answer your questions. What you'll hear: - How Nemotron Labs Diffusion approaches text generation differently from autoregressive models — and what that opens up for latency, quality, and control - How the Elastic framework produces multiple reasoning model sizes from a single training run - What's still open research and what the team is exploring next Have questions for the researchers? Bring them — this livestream is built around direct dialogue.