测试版
排行榜/Claude 3 Opus
Anthropic logo

Claude 3 Opus

来自 Anthropic · 发布于 2024-01-01

33.7
平均分
N/A
输入价格
N/A
输出价格
N/A
上下文窗口
text
类型

Tested on 8 benchmarks with 33.7% average. Top scores: MMLU (79.5%), Winogrande (77.0%), MATH level 5 (37.5%).

基准测试类别分数Bar
MMLUknowledge79.5
Winograndeknowledge77.0
MATH level 5math37.5
GPQA diamondknowledge29.6
WeirdMLcoding23.2
Cybenchcoding10.0
SimpleBenchreasoning8.2
OTIS Mock AIME 2024-2025math4.6