Skip to content
AI Primer

NVIDIA Nemotron 3 Ultra

Nemotron 3 Ultra

A specific NVIDIA Nemotron 3 model release: a 550B-total, 55B-active-parameter hybrid Mamba-Attention Mixture-of-Experts language model for reasoning and agentic workflows, released with pre-trained, post-trained, and quantized checkpoints and training datasets.

Pricing

Model profile · Current snapshot
Input / 1M
$0.60
Output / 1M
$2.75
Blended / 1M
$1.14
Output TPS
158
TTFT (s)
1.23

Model Intelligence

Context window
1,000,000 tokens
Arena ranking
38
Benchmarkable
Yes
Model level
release
Intelligence Index
38.3
Coding Index
49.3
GPQA
0.87
HLE
0.28
SciCode
0.4
IFBench
0.81
LCR
0.71
TerminalBench Hard
0.36
TAU2
0.83

Recent stories

0 linked stories
No linked stories yet.
AI PrimerAI Primer

Your daily guide to AI tools, workflows, and creative inspiration.

© 2026 AI Primer. All rights reserved.