Claude Opus 4.5 (Reasoning)
AnthropicClaude
描述
Premium model combining maximum intelligence with practical performance. Best model in the world for coding, agents, and computer use. Most robustly aligned model with best prompt injection resistance of any frontier model. Features extended thinking, 200K context window, 64K max output, and a new effort parameter for controlling reasoning depth. Pricing: $5/$25 per million tokens (input/output).
发布日期
2025-11-24
参数规模
—
上下文长度
200K
支持模态
image, pdf, text
能力雷达图
54
general
79
coding
91
reasoning
62
science估算
80
agents
80
multimodal
Science 在缺少专门科学评测时使用推理能力代理估算。
排行榜排名
基准测试分数 (LLM Stats)
Agents
OSWorld
66.3%自报
MCP Atlas
62.3%自报
Terminal-Bench 2.0
59.3%自报
Biology
GPQA
87.0%自报
Code
SWE-Bench Verified
80.9%自报
Communication
Tau2 Telecom
98.2%自报
Tau2 Retail
88.9%自报
General
MMMLU
90.8%自报
MMMU (validation)
80.7%自报
LiveBench
76.0%自报
Reasoning
ARC-AGI v2
37.6%自报
AA 评测指数
Math Index91.3
Intelligence Index40.8
Aime 250.9
Mmlu Pro0.9
Tau20.9
Livecodebench0.9
Gpqa0.9
Lcr0.7
Ifbench0.6
Scicode0.5
Terminalbench Hard0.5
Hle0.3
LLM Stats 分类评分
Language90
Physics90
Biology90
Chemistry90
Communication90
Math80
Reasoning80
Frontend Development80
General80
Healthcare80
Tool Calling80
Multimodal70
Code70
Agents60
Vision60
Spatial Reasoning40
定价
输入价格$5 / 1M tokens
输出价格$25 / 1M tokens
混合价格(3:1)$10 / 1M tokens
缓存读取价格$0.5 / 1M tokens
缓存写入价格$6.25 / 1M tokens
速度
Tokens/秒77.1
首Token延迟13.57s
首回答延迟13.57s
供应商价格排行
供应商价格排行
21 个供应商
最便宜: QiHang最贵: Cortecs
供应商输入输出
1QiHang最便宜
$0.71
$3.57
2Poe
$4.3
$21
3Jiekou.AI
$4.5
$22.5
4NanoGPT
$4.998
$25.007
5Anthropic主要
$5
$25
6302.AI
$5
$25
7Abacus
$5
$25
8Perplexity Agent
$5
$25
9OpenRouter
$5
$25
10ZenMux
$5
$25
11Kilo Gateway
$5
$25
12Cloudflare AI Gateway
$5
$25
13GitHub Copilot
$5
$25
14Vertex (Anthropic)
$5
$25
15Requesty
$5
$25
16Vercel AI Gateway
$5
$25
17LLM Gateway
$5
$25
18Vertex
$5
$25
19OrcaRouter
$5
$25
20Merge Gateway
$5
$25
21Cortecs
$5.98
$29.89
比较该模型在不同 API 供应商之间的定价。