AI BriefingNVIDIA Nemotron 3.5 Lightning added to AI Gateway for free
Key point
The open-weight reasoning model features 30B parameters with 3B active per token and a 1M token context window.
Details
NVIDIA's Nemotron 3.5 Lightning is now available for free on AI Gateway as nvidia/nemotron-3.5-lightning-free, served directly by NVIDIA. This open-weight reasoning and chat model is designed for long-running agents, sub-agent work, chatbots, and RAG systems.
Model Architecture and Capabilities
The model utilizes a hybrid Mixture-of-Experts design that interleaves Mamba-2 and MoE layers with a few attention layers. Key specifications include:
- Parameters: 30B total parameters with 3B active per token.
- Context Window: 1M tokens.
- Features: Supports tool calling and structured outputs.
Integration and Pricing
Developers can access the model via the AI SDK, Chat Completions, Responses, and Messages APIs, or through coding agents like Claude Code, Codex, and Cursor by installing the latest Vercel CLI. AI Gateway reflects provider pricing with no markup and charges no platform fee on inference, including Bring Your Own Key (BYOK) requests. Note that the provider may retain prompts for model improvement.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Amazon S3, S3 Tables, and S3 Vectors now available on Vercel Marketplace and v0