NVIDIA

NVIDIA

www.nvidia.com
+
+
+
+
Portfolio Stats
Total Models3
Multimodal0
Benchmarks Run2
Avg Performance58.6%
+
+
+
+
Latest Release
Llama-3.1 Nemotron Ultra 253B
Released: Apr 7, 2025
+
+
+
+
Release Timeline
Recent model releases by year
2025
3 models
+
+
+
+
Performance Overview
Top models and benchmark performance

Top Performing Models

By avg score
#1Llama-3.3 Nemotron Super 49B
91.3%
#2Llama 3.1 Nemotron Nano 8B
84.6%
#3Llama-3.1 Nemotron Ultra 253B
0.0%

Benchmark Categories

Coding
2
87.9%

Model Statistics

Multimodal Ratio
0%
Models with Providers
0

All Models

Complete portfolio of 3 models with advanced filtering

LicenseLinks
#01NVIDIALlama-3.1 Nemotron Ultra 253B
NVIDIA's 253B reasoning model derived from Llama 3.1 405B via neural architecture search compression
Apr 7, 2025
Apache 2.0
-
#02NVIDIALlama-3.3 Nemotron Super 49B
NVIDIA's 49B Llama 3.3-based model enhanced with Nemotron post-training for advanced reasoning
Mar 1, 2025
Apache 2.0
91.3%
#03NVIDIALlama 3.1 Nemotron Nano 8B
NVIDIA's 8B reasoning-enhanced fine-tune of Llama 3.1 for instruction-following and agentic tasks
Jan 6, 2025
Apache 2.0
84.6%
+
+
+
+
Resources
Official Website