趋近智
Claude 4.1 Opus
参数
-
上下文长度
200K
模态
Text
架构
Dense
许可证
Proprietary
发布日期
5 Aug 2025
训练数据截止日期
-
技术规格
注意力结构
Multi-Head Attention
隐藏维度大小
-
层数
-
注意力头
-
键值头
-
激活函数
-
归一化
-
位置嵌入
Absolute Position Embedding
Claude 4.1 Opus
Standard Claude 4.1 Opus without thinking mode for efficient operation. Very strong coding (76.07 LiveBench Coding) and excellent language understanding (76.75). Good web development performance (Elo 1391). Delivers high-quality results with lower latency. Well-suited for production environments requiring reliable performance.
关于 Claude 4
Anthropic's fourth generation Claude models with advanced reasoning, extended context windows up to 200K tokens, and configurable thinking effort levels. Features improved safety alignment, nuanced understanding, and sophisticated task completion. Includes Opus (most capable), Sonnet (balanced), and Haiku (fast) variants, with thinking modes that enable transparent chain-of-thought reasoning for complex problems.
其他 Claude 4 模型
评估基准
排名
#31
| 基准 | 分数 | 排名 |
|---|---|---|
Agentic Coding LiveBench Agentic | 0.53 | ⭐ 4 |
Coding LiveBench Coding | 0.76 | 10 |
Web Development WebDev Arena | 1391 | 16 |
Data Analysis LiveBench Data Analysis | 0.67 | 32 |
Mathematics LiveBench Mathematics | 0.63 | 39 |
Graduate-Level QA GPQA | 0.81 | 41 |
Reasoning LiveBench Reasoning | 0.41 | 44 |
排名
排名
#31
编程排名
#23