Claude Opus 4.6
来自 Anthropic · 发布于 2026-02-04
57.5
平均分
$5.00/1M
输入价格
$25.00/1M
输出价格
1.0M tokens (~500 books)
上下文窗口
multimodal
类型
Tested on 19 benchmarks with 57.5% average. Top scores: Chatbot Arena Elo — Coding (1542.9%), Chatbot Arena Elo — Overall (1496.6%), OTIS Mock AIME 2024-2025 (94.4%).
基准测试分数
| 基准测试 | 类别 | 分数 | Bar |
|---|---|---|---|
| Chatbot Arena Elo — Coding | arena | 1542.9 | |
| Chatbot Arena Elo — Overall | arena | 1496.6 | |
| OTIS Mock AIME 2024-2025 | math | 94.4 | |
| ARC-AGI | reasoning | 94.0 | |
| Cybench | coding | 93.0 | |
| GPQA diamond | knowledge | 87.4 | |
| SWE-Bench verified | coding | 78.7 | |
| WeirdML | coding | 77.9 | |
| Terminal Bench | coding | 74.7 | |
| ARC-AGI-2 | reasoning | 69.2 | |
| SimpleBench | reasoning | 61.1 | |
| SimpleQA Verified | knowledge | 46.5 | |
| FrontierMath-2025-02-28-Private | math | 40.7 | |
| GSO-Bench | coding | 33.3 | |
| APEX-Agents | agentic | 31.7 | |
| HLE | knowledge | 31.1 | |
| PostTrainBench | knowledge | 23.2 | |
| FrontierMath-Tier-4-2025-07-01-Private | math | 22.9 | |
| Chess Puzzles | knowledge | 17.0 |
相似模型
z-ai
57.6
Mistral AI
57.8
DeepSeek
57.9
baidu
57.0