Nemotron Super
by NVIDIA · Model · Language & code models
Nvidia Nemotron 3.5 Super VL: multimodal 120B MoE (12B active), over 600 tokens/s, open weights with training recipe, about 75 GB
Evidence
- Oct 18Less intelligent than MiniMax or Qwen 3.8 27B but much faster
- Sep 3Cited as the only other prominent fully open stack (weights plus training materials)
- Aug 16Open 30B MoE with 1M context; below Qwen 3.6, near Qwen 3.5 and Gemma 4, but fastest in its size class; FP4 about 22 GB
Cost to run
Alternatives
Built on Nemotron Super