🟢

🧪 NVIDIA Nemotron Test Lab — Standalone test page for NVIDIA's Nemotron models via integrate.api.nvidia.com · Free tier: 1,000 credits & 40 req/min · Get API key →

🔑 NVIDIA API Key

Get free API key at build.nvidia.com → · Keys auto-clear after 24 hours. 🔒 Forwarded to NVIDIA API via our proxy — never persisted on our servers.

⚡ Nemotron Synthesis

To run GPT, Claude, or Gemini alongside Nemotron, go back and enter your API keys in the Standard Lab first — they'll be picked up here automatically.

🟢 Select Nemotron Models

Select up to 4 models. MoE architecture — active params much smaller than total. Reasoning toggle: prefix your prompt with detailed thinking on to enable reasoning traces.

📋 Test Notes

  • Nemotron models use MoE architecture — only a fraction of total params are active per token
  • Reasoning: add detailed thinking on to your system prompt to enable step-by-step reasoning traces
  • API endpoint: https://integrate.api.nvidia.com/v1 (OpenAI-compatible format)
  • Free tier: ~1,000 credits on signup, 40 req/min. Errors here don't affect the main lab
  • Ultra (500B) model not yet available — will be added when released later in 2026