Skip to main content
leaderboard
.cn
Toggle Sidebar
Search models or organizations…
⌘K
Explore
Frontier Models
Historical Trends
Local Models
Compare Models
Ranking Studio
Library
Models
Pricing
Organizations
Benchmarks
Data sources
Settings
Toggle Sidebar
leaderboard
.cn
Loading page content
Feedback
Loading page content
Back to models
Nemotron-3-Ultra
Open
Share
Add to comparison
Tool use · Reasoning
Overview
Deployment
Evaluations
Evaluation overview
Overall rank
#77
72.8 points · 202 models
Open-weight rank
#26
75.9 points · 120 models
Data coverage
41 items
2 sources · updated through 2026-09-01
Domain performance
Coding
#69 / 182
71.4
Research
#69 / 196
83.7
Agents
#86 / 148
65.9
Long context
#52 / 95
75.8
Evaluation highlights
All evaluations
MultiChallenge
Reasoning · Nemotron-3-Ultra Technical Report
63.8%
3
/ 22
RULER (1M)
Long context · Nemotron-3-Ultra Technical Report
94.7%
1
/ 4
PinchBench
Agents · Nemotron-3-Ultra Technical Report
90%
3
/ 8
IOI 2025
Coding · Nemotron-3-Ultra Technical Report
570pts
3
/ 5
OmniScience Non-Hallucination
Reasoning · Nemotron-3-Ultra Technical Report
78.7%
1
/ 7
MMLU-Pro
Reasoning · Nemotron-3-Ultra Technical Report
86.8%
17
/ 96
Basic information
Context (input / output)
1.05M / 16K
Parameters
550B / 55B active
Release date
2026-06
Input modalities
Text
License
OpenMDW License Agreement v1.1
Output modalities
Text
Architecture
Mamba-Transformer Hybrid MoE
Availability
vLLM · BF16 · 256K
Deployment options
BF16 · 16 × 141 GB VRAM
Related links
Official site
HuggingFace
ModelScope