跳转到主要内容

Claude Opus 5.5 (Max, Default Fallback)

AnthropicClaudeProprietary

描述

Claude Opus 5.5 is the first model in Anthropic's Claude 5.5 family. It performs at Claude Fable 5.1 level on most work at about 40% lower cost than Claude Opus 5, and targets long-running agentic coding and knowledge work. It accepts text and image input with text output, supports multilingual and vision workloads, tool use, and adaptive thinking that is always on (API default effort medium). Context window is 1M tokens with 128K max output on the sync Messages API; reliable knowledge and training-data cutoffs are June 2026. First-party pricing is $4/$20 per million input/output tokens, with cache reads at $0.20 and 5-minute cache writes at $5 per million tokens. Fast mode is available in Claude Code and on the Claude Platform at $8/$40 per million input/output tokens with up to 2.5x speed (same model id; not a separate deployment). Capability scores are self-reported from the Claude Opus 5.5 System Card (Table 8.1.A and §8) and the launch announcement. Available on the Claude API as `claude-opus-5-5`.

发布日期
2026-09-22
参数规模
—
上下文长度
1.0M
支持模态
image, pdf, text

能力雷达图

59
general
70
coding
70
reasoning
60
science估算
60
agents
80
multimodal

缺少专门科学评测时,Science 由 LLM Stats 科学得分或推理能力估算。

排行榜排名

领域#排名分数来源
智能体能力模型榜4
72.0
LS
代码能力榜3
96.0
LB
通用能力榜3
91.0
AA
数学推理2
97.0
LB
多模态榜19
63.0
LS
推理能力5
92.0
LB
科学能力2
89.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

AA-Briefcase v1.11822.00 / 3000自报
Program Bench91.2%自报
OfficeQA78.9%自报
Toolathlon-Verified77.8%自报
OfficeQA Pro67.7%自报
AutomationBench40.0%自报

Biology

BioMysteryBench89.3%自报
LatchBio SpatialBench Verified72.0%自报

Code

BenchCAD (with Python tool)96.2%自报
DeepSWE 1.174.2%自报
BenchCAD73.0%自报
FrontierSWE V262.3%自报
CursorBench 4.057.8%自报

General

GDPval-AA 2.11846.00 / 3000自报
Global-MMLU94.3%自报

Healthcare

HealthBench Professional65.6%自报
HealthBench60.6%自报
LifeSci Diamond (2026-11)3.6%

Language

MILU93.1%自报

Legal

Legal Agent Benchmark8.3%自报

Math

ArXivMath91.2%自报

Multimodal

OSWorld 2.081.8%自报

Reasoning

SWE-bench Multilingual93.9%自报
SWE-Bench ProPrinceton NLP (2024)89.9%自报
Humanity's Last Exam (with tools, text-only)67.7%自报
Terminal-Bench 4.066.4%自报
Humanity's Last Exam (no tools, text-only)64.4%自报
SWE-Bench Multimodal61.4%自报
Terminal-Bench-Science 0.158.7%自报
FrontierCode 1.154.4%自报

Science

LatchBio SingleCellBench61.2%自报

Vision

Chartography89.0%自报

AA 评测指数

(Artificial Analysis)
Engineering(Artificial Analysis)
65.3
Strategy And Ops
61.4
Economics
60.3
Finance And Accounting
59.4
Intelligence Index(Artificial Analysis)
59.1
Legal
58.0
Healthcare And Medical
57.9

LLM Stats 分类评分

(LLM Stats (zeroeval))
Productivity
100
Agents
100
Reasoning
100
General
100
Language
90
Multimodal
80
Vision
80
Math
70
Science
70
Biology
70
Code
70
Healthcare
60
Tool Calling
60
Legal
10

定价

输入价格$4 / 1M tokens
输出价格$20 / 1M tokens
混合价格(3:1)$8 / 1M tokens
缓存读取价格$0.2 / 1M tokens
缓存写入价格$5 / 1M tokens

速度

Tokens/秒98.3
首Token延迟765.80s
首回答延迟765.80s

供应商价格排行

供应商价格排行

1 个供应商

供应商输入输出
1Anthropic主要
$4
$20

比较该模型在不同 API 供应商之间的定价。

外部链接