Nemotron-3

also referred to as: nemotron 3

part of Nemotron includes Nemotron 3 Super

3 statements across 1 episodes · 1 bullish · 0 bearish · 1 people on the record · first statement Jul 2, 2026 by Bryan Catanzaro · said 11 times in 1 episodes since 2026 · across every show →

Mentions by year, the whole family

brought up most by Matt Turck (6), Bryan Catanzaro (5)

mentionsepisodesper episode tap a year for its mentions

2026 11 mentions in 1 episode

Inside Nemotron & NVIDIA’s AI Lab | Bryan Catanzaro Jul 2, 2026 · 11 mentions · from 27:34
every mention in 2026, scene by scene →

every mention, scene by scene, with the transcript →

Everything said about Nemotron-3, oldest first

Jul 2, 2026 positive

Assertion Supported

Catanzaro: Nemotron 3's Latent MoE quadruples experts for same inference cost

“Latent MOE is a specific innovation that we have in NemoTron three family. And what it does is actually reduces the amount of communication that has to be sent through NVLink during MOE computations by basically down projecting it. So, you know, every token pr…”

Bryan Catanzaro Jul 2, 2026 ▶ 46:00 Inside Nemotron & NVIDIA’s AI Lab | Bryan Catanzaro

Jul 2, 2026 neutral

Assertion Supported

NVIDIA details Nemotron 3 parameter specs from 3B to 55B active

“Nano is a thirty billion total three billion active parameter model. Super is one 20 and 12, and Ultra is five 50 and 55.”

Bryan Catanzaro Jul 2, 2026 ▶ 33:50 Inside Nemotron & NVIDIA’s AI Lab | Bryan Catanzaro

Jul 2, 2026

Assertion Supported

Catanzaro: NVIDIA used 10 to 15 teacher models for Nemotron-3 distillation

“There's with NemoTron three, I think we had about 10, 10 or 15 of these teachers.”

Bryan Catanzaro Jul 2, 2026 ▶ 53:24 Inside Nemotron & NVIDIA’s AI Lab | Bryan Catanzaro

Nemotron-3

Mentions by year, the whole family

Everything said about Nemotron-3, oldest first

Turn any episode into a week of clips.