Command Palette
Search for a command to run...

Nvidia Unveils Nemotron 3 Ultra, a 550B-Parameter Open-Weights AI Model Due This Week

aiai-modelingai-model-releasesai-open-models 7 posts · 6 accounts

Nvidia said Nemotron 3 Ultra, a 550-billion-parameter open-weights AI model with 55 billion active parameters, will be released this week after Jensen Huang announced it during his Computex keynote. Artificial Analysis, which said it evaluated the model with Nvidia, gave it a score of 48 on its Intelligence Index, ahead of Gemma 4 31B at 39, Nemotron 3 Super at 36 and gpt-oss-120b at 33.

Artificial Analysis said Nemotron 3 Ultra ran at more than 300 tokens per second on a pre-release DeepInfra endpoint and called it the largest Nemotron 3 model so far, with roughly 550 billion total parameters and 90% sparsity. The firm said that makes it the strongest US open-weights model on its benchmark, though still behind Moonshot's Kimi K2.6, which scored 54.

From the sources (7 posts)

@mweinbach

Nemotron 3 Ultra model! 550B parameter, near state of the art open weight (on par with GLM-5.1 and Kimi K2.6) We shall see how it is in practice

@artificialanlys

NVIDIA just announced the release of Nemotron 3 Ultra in Jensen Huang's Computex keynote: at 550B parameters (55B active), this is the largest Nemotron 3 model to date, and it is the most intelligent US open weights model We partnered with

@cointelegraph

🔥 JUST IN: Nvidia CEO Jensen Huang announces the launch of its new AI model, Nemotron 3 Ultra during Computex 2026.

@nvidiaai

Nemotron 3 Ultra is coming this week. ⌛️

@ctnzr

Nemotron 3 Ultra: Frontier smart. 5X faster. 30% cheaper. 💚💚💚

@eliebakouch

this plot goes very hard

@ctnzr

RT @NVIDIAAI: Nemotron 3 Ultra is coming this week. ⌛️

Preview built on a synthetic news corpus (16 weeks, Apr–Jul 2026). Impact calls are model reads, not price data.

About Archive