NVIDIA releases Nemotron 3.5 Super VL: 120B total and 12B active parameters, combining reasoning and tool use with image, document and video understanding; NVFP4 weights are available.

NVIDIA’s published model-speed chart using Artificial Analysis scores and 100K input tokens; its highlighted speed comparison is with Inkling Small.

Published 16 Oct 2026 · 01:29 UTC

NVIDIA Developer ↗ · Source dated 15 Oct 2026

Browse all Live updates · Read Features