Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)
NVIDIALlamaOpen WeightLlama 3.1 Community License
描述
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.
發布日期
2025-04-07
參數規模
253.0B
上下文長度
—
支援模態
—
能力雷達圖
34
general
33
coding
72
reasoning
47
science估算
70
agents
0
multimodal
Science 在缺少專門科學評測時使用推理能力代理估算。
排行榜排名
基準測試分數 (LLM Stats)
Biology
GPQA
76.0%自報
Code
LiveCodeBench
66.3%自報
General
IFEval
89.5%自報
BFCL v2
74.1%自報
Math
MATH-500
97.0%自報
AIME 2025
72.5%自報
AA 評測指數
Math Index63.7
Intelligence Index15.0
Coding Index13.1
Math 5001.0
Mmlu Pro0.8
Aime0.7
Gpqa0.7
Livecodebench0.6
Aime 250.6
Ifbench0.4
Scicode0.3
Tau20.1
Hle0.1
Lcr0.1
Terminalbench Hard0.0
LLM Stats 分類評分
Structured Output90
Instruction Following90
Biology80
Chemistry80
General80
Math80
Physics80
Reasoning80
Tool Calling70
Code70
定價
輸入價格$0.6 / 1M tokens
輸出價格$1.8 / 1M tokens
混合價格(3:1)$0.9 / 1M tokens
速度
Tokens/秒40.9 tokens/s
首Token延遲0.75s
首回答延遲49.64s
可用提供商
(LS 內部計價單位)暫無提供商資料