The leading open models built by the community, optimized and accelerated by NVIDIA's enterprise-ready inference runtime.

• Moonshotai

  kimi-k3

  ~2.8T hybrid KDA+MLA multimodal MoE for long-horizon coding, agentic tool use, and image understanding.
• DeepSeek AI

  deepseek-v4-pro-0813

  DeepSeek V4 scales to 262K-token context windows with efficient MoE architecture for coding tasks.
• NVIDIA

  nemotron-3.5-lightning-30b-a3b

  Fastest 30B A3B MoE model with leading domain accuracy for specialized agentic tasks
• NVIDIA

  nemotron-3-ultra-550b-a55b

  Open, efficient hybrid Mamba-Transformer MoE with 1M context, excelling in agentic reasoning, coding, planning, tool calling, and more

Get started with workflows and code samples to build AI applications from the ground up.

• NVIDIA

  NemoClaw for Hermes Agent

  Deploy Hermes agents that learn from team workflows, create reusable skills, and get better with every interaction.
• NVIDIA

  NVIDIA AI-Q Blueprint for intelligent agents

  AI agents that connect, retrieve, and reason on enterprise data—making information accessible, actionable, and intelligent.
• NVIDIA

  Build a Video Search and Summarization (VSS) Agent

  Ingest massive volumes of live or archived videos and extract insights for summarization and interactive Q&A
• NVIDIA

  Cosmos Dataset Search

  Accelerate post-training of end-to-end autonomous vehicle stacks with vector search and retrieval for large video datasets.