Nemotron 3 Super (free) by OpenRouter

$10 off the fastest Kimi K3 and GLM 5.2$10 off the fastest Kimi K3 and GLM 5.2

Model details

Nemotron 3 Super (free)

Nemotron 3 Super free is NVIDIA's open hybrid Mamba-Transformer Mixture-of-Experts model, pairing a 120B-parameter backbone with just 12B active parameters per forward pass. That sparse design is positioned by NVIDIA as a compute-efficient path for complex multi-agent workflows, and the model is published under the NVIDIA Open License with full weights, datasets, and recipes available. Reinforcement learning across more than ten environments is cited as the training foundation behind results on reasoning and agent benchmarks such as AIME 2025, TerminalBench, and SWE-Bench Verified.

In practical use, the model is a strong fit for long-context text tasks that benefit from reasoning plus structured tool interaction, including deep document analysis, tool-augmented pipelines, and multi-step agent loops. Independent index readings of roughly 25–26 on intelligence, 37–38 on coding, and 8–9 on agentic tasks place it as a capable generalist rather than a top-of-leaderboard specialist. The combination of a large context budget, reasoning support, tool calling, and structured output makes it well suited for experimentation and production prototypes where free access and open weights matter more than raw benchmark dominance.

OpenRouternvidia/nemotron-3-super-120b-a12b:freenemotron

Open provider page Provider docs

Quick Info

Provider: OpenRouter
Model key: nvidia/nemotron-3-super-120b-a12b:free
Release date: Mar 11, 2026
Last updated: Mar 11, 2026
Input modalities
Output modalities
Capabilities

Cost

A provider subscription or plan supersedes token-based pricing for this model.

Limits

Output tokens: 235,929 tokens
Context window: 262,144 tokens

Latest news about Nemotron 3 Super (free)

OpenRouter

Official sourceComparison

The OpenRouter compare page for Nemotron 3 Super (free) wires the model into comparison views against flagship models such as Claude Fable 5 (batch), GPT-5.5 (batch), and Gemini 3.1 Pro Preview (batch), as well as affordable picks including DeepSeek V4 Flash 0423, Hy3 preview, and Gemini 2.5 Flash Lite (batch). It also Technically, the comparison landing page offers no new specs beyond the model page and is mostly scaffolding for side-by-side benchmarking, pricing, and context-length comparisons. The excerpt is heavily UI chrome and competitive-group labels, so fresh technical signal is limited, but it confirms Nemotron 3 Super (free

OpenRouter

Official sourceBenchmark

OpenRouter's main model listing for NVIDIA Nemotron 3 Super (free) repeats the same architectural profile: a 120B-parameter open hybrid Mamba-Transformer MoE with 12B active parameters, Multi-Token Prediction, a 1M-token advertised context window, and NVIDIA Open License weights. It highlights the same headline benchma The non-API listing reports a live routing snapshot that differs slightly from the API page: P50 latency around 1.36 seconds, throughput near 35 tokens per second, and uptime of about 93.01%, illustrating that these numbers fluctuate rather than constituting a fixed SLA. As with the API page, it reiterates that the fre

OpenRouter

Official sourceOfficial

Effective pricing across providers for NVIDIA: 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. Built on a hybrid Mamba-Transformer Mixture-of-Experts architecture