Model details
NVIDIA Nemotron 3 Super is a hybrid Mamba-Transformer model engineered to handle the demands of complex, agentic AI systems. By utilizing a Mixture-of-Experts architecture with 120 billion total parameters and 12 billion active parameters per forward pass, the model balances high-level reasoning with significant compute efficiency. It incorporates LatentMoE for improved accuracy and MTP layers that enable native speculative decoding, allowing it to maintain high performance even as multi-agent workflows generate large volumes of tokens.
The model was developed through a rigorous training pipeline that included pre-training on 25 trillion tokens, followed by supervised fine-tuning and reinforcement learning. This process, combined with its ability to support a 1 million token context length, makes it well-suited for tasks ranging from deep literature search and molecular understanding to automated software development. By delivering high throughput and accuracy, the model provides a scalable foundation for enterprise applications that require reliable performance in data-intensive environments.
Transparent token rates
Rates are shown per one million tokens. Combined means one million input plus one million output tokens.
Kilo Gateway
NVIDIA Corporation (NASDAQ:NVDA) is one of the most profitable blue chip stocks to invest in now. NVIDIA Corporation (NASDAQ:NVDA) announced on March 11 the launch of the NVIDIA Nemotron 3 Super, which is a 120‑billion‑parameter open model with 12 billion active parameters designed to run complex agentic AI systems at
Kilo Gateway
OCI Generative AI continues to expand how developers and enterprises work with foundation models. Today, we’re announcing support for importing open-weights model NVIDIA Nemotron 3 Super, enabled by a new Model Import capability in OCI Generative AI.
Kilo Gateway
Nemotron 3 Super is now available as a fully managed and serverless model on Amazon Bedrock, joining the Nemotron Nano models that are already available...
Kilo Gateway
NVIDIA has unveiled Nemotron 3 Super, the next open AI model in its Nemotron series, and of course, the whole thing is being marketed as a major breakthrough…
Kilo Gateway
NVIDIA's Nemotron 3 Super 120B A12B is available via OpenRouter as a free endpoint under the model key nvidia/nemotron-3-super-120b-a12b:free, released on March 11, 2026. The supplied OpenRouter excerpt describes it as a 120B-parameter open hybrid MoE model that activates only 12B parameters per pass, built on a Mamba- For developers evaluating or routing to Nemotron 3 Super, the OpenRouter page positions it as purpose-built for complex multi-agent applications with a 1M-token native context window for long-term coherence, cross-document reasoning, and multi-step task planning, along with strong benchmark performance on AIME 2025, Te
This exact model name is also listed by 7 other providers.
