Nvidia Nemotron 3 Super 120B Thinking
Nemotron 3 Super is the middle model of NVIDIA's open Nemotron 3 family, an MoE with 120B total and 12B active parameters. It pairs Mamba-2 with attention layers, works in text, and has switchable reasoning for agents, tool use and RAG.
- Price per 1M tokens
- $0.075 / $0.38
- input / output
- Cached input
- $0.075
- per 1M tokens
- Context
- 262K tokens
- 16K max output
- Providers
- 2
Providers
Compare all 2 providers of Nvidia Nemotron 3 Super 120B Thinking by price, speed and uptime, and call any of them through one OpenAI-compatible API.
Provider prices are what each provider charges. You pay the price above until pass-through billing starts; then you pay the provider's price directly.
262K context · 16K max output
- 1 more provider through partner hosts · from $0.06 in · $0.26 out
Per 1M tokens: the provider's own price. Speed and uptime show here once measured. “Requests go here” marks where a request that names no provider goes now. Logos are trademarks of their owners.
* No cache discount from this provider: cached tokens cost the input price.
Price history
API
curl https://api.opencheap.si/v1/chat/completions \
-H "Authorization: Bearer $OPENCHEAP_API_KEY" -H "Content-Type: application/json" \
-d '{ "model": "opencheap/nemotron-3-super-120b-a12b:thinking", "messages": [{ "role": "user", "content": "Hi" }] }'