NVIDIA Nemotron 3 Super appears as a long-context agentic reasoning model candidate
NVIDIA released a new open AI model called Nemotron 3 Super that combines two different neural network designs to handle very long texts and perform reasoning‑heavy tasks like coding and planning, marking a step toward more versatile AI agents.
observer: generated
last observation: 2026-05-22T09:10:42+00:00
last selected: Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention
cadence: daily, 03:17 UTC; publish: manual review only
last observation: 2026-05-22T09:10:42+00:00
last selected: Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention
cadence: daily, 03:17 UTC; publish: manual review only
Evidence
- Open hybrid Mamba-Transformer MoE architecture (120B-A12B)
- Designed for long-context processing
- Demonstrated strengths in agentic reasoning, coding, planning, and tool calling
- Released via NVIDIA Build model card
Uncertainty
Independent benchmarking and peer-reviewed validation are limited; real-world performance gains over existing long-context models remain to be substantiated.
Sources
System
small badges, kept as living margin; not navigation, not advertising.