Nvidia drops open 30B MoE model and agentic router
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Nvidia released Nemotron 3.5 Lightning, an open 30B-parameter Mixture-of-Experts (MoE) model
- Nvidia claims Nemotron 3.5 Lightning delivers up to 4x faster output than prior versions
- Nvidia also launched NeMo Switchyard, an agentic AI model router for orchestrating model calls
- SiliconANGLE's Kyt Dotson reports both were unveiled together as two new services from Nvidia Corp.
- Hacker News also surfaced the Nemotron 3.5 Lightning release to its community alongside SiliconANGLE's coverage
Why it matters: Nvidia pairs an open 30B-parameter MoE model claiming 4x faster output with an agentic AI model router, giving developers building agentic AI systems a single-vendor path to high-speed inference and model orchestration in one release.
Ask SkimNews
