NVIDIA Launches Nemotron 3.5 Lightning: 30B MoE Model for High-Throughput Agents
baseten · x · 2026-08-12
NVIDIA has released Nemotron 3.5 Lightning, a new open-weight large language model. The model features a 30B parameter Mixture-of-Experts (MoE) architecture with only 3B active parameters and supports a massive 1M context window.
Officially built for high-throughput AI agent workflows, the model distills a larger foundation model into a more efficient architecture, achieving up to 4x the throughput of comparable open models. This speed allows agents to perform more reasoning iterations and deeper validation, bypassing traditional speed-accuracy trade-offs in complex tasks ranging from personal assistants to cybersecurity and financial services. It is now available for deployment on the Baseten platform.
Related event: NVIDIA Unveils Nemotron 3.5 Lightning and NeMo Switchyard Router(55 posts)→
More from coding & agent
- Dev uses Grok to clean up 3D printer firmware, catches it sneaking in nonsense code — yacineMTB · 2026-10-16
- Guardrails can make risky delegation work — but it takes a full-time team, says iandanforth — iandanforth · 2026-10-16
- What can you truly delegate? Only tasks with unambiguous, verifiable results — iandanforth · 2026-10-16
- Yacine: vibe-coded repos with 3,330 agent-written tests are all useless — yacineMTB · 2026-10-16
- Dev argues coding models now beat nearly all engineers in every branch of software — zack_overflow · 2026-10-16
- Q4-quantized coder model on 8-core CPU builds interactive solar system demo in one shot — IntrepidMindExplorer · 2026-10-16