Nemotron 3 Super 120B A12B

OpenRouter's comparison between MiniMax M3 and Nemotron 3 Super 120B A12B shows both models sitting in the ~1M-token context tier, with MiniMax M3 offering 1,048,576 tokens at $0.23 input and $0.96 output per million tokens versus Nemotron 3 Super's 1,000,000-token window at $0.085 input and $0.40 output. Either can be The page reinforces that Nemotron 3 Super is among the cheapest 1M-context listings on OpenRouter, undercutting MiniMax M3 by roughly 2.7x on input and 2.4x on output pricing while matching context scale. As with the other auto-generated compare page, no benchmark or quality metrics are present in the excerpt, so the v

OpenRouter's head-to-head comparison of Cohere's Command A and NVIDIA's Nemotron 3 Super 120B A12B positions the two models for developers choosing a 1M-context-capable alternative at very different price points. Command A offers a 256,000-token context window at $2.50 per million input tokens and $10 per million outpu The comparison emphasizes Nemotron 3 Super's roughly 29x lower input cost and 25x lower output cost against Command A, alongside a 4x larger effective context window for long-document workloads. There are no benchmark or quality scores in the supplied excerpt, so the page is most useful as a pricing and context sizing

OpenRouter hosts a free endpoint for NVIDIA's Nemotron 3 Super 120B A12B, a 120B-parameter open hybrid MoE model that activates only 12B parameters per forward pass, using a Mamba-Transformer architecture with Multi-Token Prediction. The listing confirms a 1M-token context window for long-horizon agent coherence, cross On this free endpoint, modalities are text, pricing is Free, and the listed context is 262K, with routing through NVIDIA as the sole provider showing P50 latency of 1.11s, 44 tps throughput, and 93.83% uptime. Developers should note the data-consent clause: free-tier usage is logged by NVIDIA per the NVIDIA API Trial T