Skip to main content

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)

NVIDIALlamaOpen WeightLlama 3.1 Community License

Description

A 253B parameter derivative of Meta Llama 3.1 405B Instruct, developed by NVIDIA using Neural Architecture Search (NAS) and vertical compression. It underwent multi-phase post-training (SFT for Math, Code, Reasoning, Chat, Tool Calling; RL with GRPO) to enhance reasoning and instruction-following. Optimized for accuracy/efficiency tradeoff on NVIDIA GPUs. Supports 128k context.

Release Date
2025-04-07
Parameters
253.0B
Context Length
Modalities

Capability Radar

31
general
58
coding
72
reasoning
47
scienceest.
70
agents
0
multimodal

Science uses a reasoning proxy when dedicated science benchmarks are unavailable.

Rankings

Domain#RankScoreSource
Code Ranking307
28.0
AA
General Ranking314
34.0
AA
Math Reasoning108
73.0
AA
Science192
49.0
AA

Benchmark Scores (LLM Stats)

Biology

GPQA76.0%SR

Code

LiveCodeBench66.3%SR

General

IFEval89.5%SR
BFCL v274.1%SR

Math

MATH-50097.0%SR
AIME 202572.5%SR

AA Evaluation Indices

Math Index
63.7
Intelligence Index
9.1
Math 500
1.0
Mmlu Pro
0.8
Aime
0.7
Gpqa
0.7
Livecodebench
0.6
Aime 25
0.6
Ifbench
0.4
Scicode
0.3
Tau2
0.1
Hle
0.1
Lcr
0.1
Terminalbench Hard
0.0

LLM Stats Category Scores

Instruction Following
90
Structured Output
90
Math
80
Physics
80
Reasoning
80
General
80
Biology
80
Chemistry
80
Code
70
Tool Calling
70

Pricing

Input Price$0.6 / 1M tokens
Output Price$1.8 / 1M tokens
Blended Price (3:1)$0.9 / 1M tokens

Speed

Tokens/sec52.2
Time to First Token0.70s
Time to Answer39.03s

Provider Price Ranking

Provider Price Ranking

3 providers

Cheapest: NVIDIAMost Expensive: LLM Gateway
ProviderInputOutput
1NVIDIAPRIMARY
$0.6
$1.8
2Nebius Token Factory
$0.6
$1.8
3LLM Gateway
$0.6
$1.8

Compare pricing across different API providers for this model.

External Sources