Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)

NVIDIAUnited States

A large-scale reasoning model from NVIDIA's Nemotron family, built upon the Llama 3.1 architecture. It is optimized for complex, multi-step reasoning tasks and is designed to deliver high accuracy in logical inference and problem-solving.

ReasoningCoding

Input / 1M tokens

$0.60

Output / 1M tokens

$1.80

Output tokens/s

52.03

First-token seconds

0.72s

Supported plans

Benchmark history

Evaluations

TAU2

Measured Jul 10, 2026Source

Score

0.11

Terminalbench Hard

Measured Jul 10, 2026Source

Score

0.02

Lcr

Measured Jul 10, 2026Source

Score

0.07

Ifbench

Measured Jul 10, 2026Source

Score

0.38

Aime 25

Measured Jul 10, 2026Source

Score

0.64

Aime

Measured Jul 10, 2026Source

Score

0.75

Math 500

Measured Jul 10, 2026Source

Score

0.95

Scicode

Measured Jul 10, 2026Source

Score

0.35

Livecodebench

Measured Jul 10, 2026Source

Score

0.64

Hle

Measured Jul 10, 2026Source

Score

0.08

Gpqa

Measured Jul 10, 2026Source

Score

0.73

Mmlu Pro

Measured Jul 10, 2026Source

Score

0.83

Artificial Analysis Math Index

Measured Jul 10, 2026Source

Score

63.7

Artificial Analysis Intelligence Index

Measured Jul 10, 2026Source

Score

9.1

Plan availability

Products and plans that support this model

No products or plans have been linked to this model yet.

User ratings

Loading ratings...

Discussion

Thinking... Make sure you are connected to GitHub server