Nemotron 3 Super
NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications. Built on a hybrid Mamba-Transformer Mixture-of-Experts architecture with multi-token prediction (MTP), it delivers over 50% higher token generation compared to leading open models. The model features a 1M token context window for long-term agent coherence, cross-document reasoning, and multi-step task planning. Latent MoE enables calling 4 experts for the inference cost of only one, improving intelligence and generalization. Multi-environment RL training across 10+ environments delivers leading accuracy on benchmarks including AIME 2025, TerminalBench, and SWE-Bench Verified. Fully open with weights, datasets, and recipes under the NVIDIA Open License, Nemotron 3 Super allows easy customization and secure deployment anywhere — from workstation to cloud.
nvidia/nemotron-3-super-120b-a12b
- Context
- 1M tokens
- Completion cap
- 16,384
- Tools
- Yes
- JSON
- Yes
- Released
- 2026-03-11
Call it from Geek Hub
Same OpenAI SDK. Change the base URL and the model id.
curl https://api.geekhub.mx/v1/chat/completions \
-H "Authorization: Bearer $GEEKHUB_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "nvidia/nemotron-3-super-120b-a12b",
"messages": [{"role": "user", "content": "Hola"}]
}'Get an API keyFrequently asked questions
- What is Nemotron 3 Super?
- NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications. Built on a hybrid Mamba-Transformer Mixture-of-Experts architecture with multi-token prediction (MTP), it delivers over 50% higher token generation compared to leading open models. The model features a 1M token context window for long-term agent coherence, cross-document reasoning, and multi-step task planning. Latent MoE enables calling 4 experts for the inference cost of only one, improving intelligence and generalization. Multi-environment RL training across 10+ environments delivers leading accuracy on benchmarks including AIME 2025, TerminalBench, and SWE-Bench Verified. Fully open with weights, datasets, and recipes under the NVIDIA Open License, Nemotron 3 Super allows easy customization and secure deployment anywhere — from workstation to cloud. Nemotron 3 Super runs on the Geek Hub API (OpenAI-compatible). Model id: nvidia/nemotron-3-super-120b-a12b.
- Is Nemotron 3 Super free?
- No. Input is $0.0927 per 1M tokens and output is $0.436 per 1M tokens on Geek Hub (markup included).
- What is the context length of Nemotron 3 Super?
- Nemotron 3 Super has a 1M tokens context window. It supports up to 16,384 completion tokens.
- Does Nemotron 3 Super support tool calling and structured outputs?
- Nemotron 3 Super accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.
- When was Nemotron 3 Super released?
- Nemotron 3 Super was released on 2026-03-11.
More models from NVIDIA
- Nemotron 3 Nano 30B A3BNVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully
- Nemotron 3 Nano 30B A3B (free)NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully
- Nemotron 3 Nano Omni (free)NVIDIA Nemotron™ 3 Nano Omni is a 30B-A3B open multimodal model designed to function as a perception and context sub-agent in enterprise agent systems. It accepts text, image, vide
- Nemotron 3 Super (free)NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications. B
- Nemotron 3 UltraNVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba m
- Nemotron 3 Ultra (batch)NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba m