Claude Haiku 5.5 vs. Claude Opus 5.5：发布比较

Claude Haiku 5.5 与 Claude Opus 5.5 发布比较：分别包含 5 个和 5 个模型，各自具有不同的智能、性能与价格特征。下方对全部 10 个模型的关键指标进行了比较。

智能方面，Claude Opus 5.5 得分最高：Claude Opus 5.5 (Max, Default Fallback)（58），Claude Haiku 5.5 为 Claude Haiku 5.5 (Max)（36）。
输出速度方面，Claude Haiku 5.5 最快：Claude Haiku 5.5 (Max)（241 token/秒），Claude Opus 5.5 为 Claude Opus 5.5 (Max, Default Fallback)（98 token/秒）。
单任务成本方面，Claude Haiku 5.5 最低：Claude Haiku 5.5 (Low)（$0.04），Claude Opus 5.5 为 Claude Opus 5.5 (Low, Default Fallback)（$0.98）。

比较的发布


	Claude Opus 5.5 (max with fallback)	Claude Opus 5.5 (xhigh with fallback)	Claude Opus 5.5 (high with fallback)	Claude Opus 5.5 (medium with fallback)	Claude Haiku 5.5 (max)	Claude Opus 5.5 (low with fallback)	Claude Haiku 5.5 (xhigh)	Claude Haiku 5.5 (high)	Claude Haiku 5.5 (medium)	Claude Haiku 5.5 (low)
Intelligence Index	58	57	52	48	36	35	34	30	25	19
财务与会计	59	56	53	51	44	42	41	38	35	29
战略与运营	61	59	58	56	39	48	39	37	33	27
法律	58	55	52	50	39	43	37	35	33	27
医疗与健康	58	56	52	51	37	45	34	32	31	25
工程	65	62	58	54	42	43	40	36	32	27
经济学	62	60	57	54	44	45	43	39	36	30
AA-Briefcase v1.1	1800	1760	1681	1621	1575	1278	1524	1440	1369	1114
AutomationBench-AA	69.5%	65.0%	63.2%	61.2%	35.4%	52.9%	36.0%	33.7%	28.6%	22.9%
Terminal-Bench 4.0	59.6%	59.6%	56.6%	52.5%	32.8%	31.3%	29.3%	21.7%	15.2%	12.6%
FrontierCode 2.0	46.4%	43.8%	35.9%	33.1%	34.9%	15.8%	29.9%	23.2%	16.1%	8.0%
AA-Omniscience v1.1	48	43	42	40	12	39	8	5	4	2
GDP.pdf	26.2%	26.6%	28.8%	25.6%	20.8%	25.6%	18.2%	17.2%	15.2%	11.2%
HLE-Diamond	67.9%	64.0%	55.6%	51.4%	29.6%	41.7%	28.8%	23.7%	16.8%	14.1%
CritPt v2 (Hard Subset)	61.0%	63.0%	58.0%	52.5%	41.0%	33.5%	44.5%	42.5%	30.0%	21.5%
Terminal-Bench-Science 0.1	59.0%	61.9%	49.0%	43.3%	20.0%	24.3%	18.1%	10.0%	7.1%	1.9%

每任务成本美元	$10.04	$6.05	$3.46	$2.45	$0.41	$0.98	$0.25	$0.16	$0.09	$0.04
输入价格美元/100 万 Token	$4.00	$4.00	$4.00	$4.00	$0.10	$4.00	$0.10	$0.10	$0.10	$0.10
输出价格美元/100 万 Token	$20.00	$20.00	$20.00	$20.00	$0.50	$20.00	$0.50	$0.50	$0.50	$0.50
缓存命中价格美元/100 万 Token	$0.20	$0.20	$0.20	$0.20	$0.01	$0.20	$0.01	$0.01	$0.01	$0.01
缓存写入价格美元/100 万 Token	$5.00	$5.00	$5.00	$5.00	$0.125	$5.00	$0.125	$0.125	$0.125	$0.125

输出速度Token/秒	98	84	81	82	241	82	196	187	175	185
首 Token 延迟秒	700.16	119.09	33.57	16.48	310.72	5.24	67.20	24.31	10.36	7.51
首个回答 Token 时间秒	700.16	119.09	33.57	16.48	310.72	5.24	67.20	24.31	10.36	7.51
端到端响应时间秒	705.25	125.03	39.78	22.59	312.80	11.33	69.75	26.99	13.21	10.22
每任务耗时秒	1,132.59	778.91	463.62	336.65	593.47	146.94	441.28	298.26	192.95	97.99

上下文窗口	1M	1M	1M	1M	1M	1M	1M	1M	1M	1M
开发商	Anthropic	Anthropic	Anthropic	Anthropic	Anthropic	Anthropic	Anthropic	Anthropic	Anthropic	Anthropic
许可证	专有	专有	专有	专有	专有	专有	专有	专有	专有	专有
输入模态	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像	支持：文本和图像
输出模态	支持：文本	支持：文本	支持：文本	支持：文本	支持：文本	支持：文本	支持：文本	支持：文本	支持：文本	支持：文本
服务商

Updated

智能

Artificial Analysis Intelligence Index · Higher is better

速度

Output tokens per second · Higher is better

每任务成本

Weighted average cost (USD) per Intelligence Index task · Lower is better

智能Updated

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v5.0 incorporates 9 evaluations: AA-Briefcase v1.1, AutomationBench-AA, Terminal-Bench 4.0, FrontierCode 2.0, AA-Omniscience v1.1, GDP.pdf, HLE-Diamond, CritPt v2 (Hard Subset), and Terminal-Bench-Science 0.1

Artificial Analysis Intelligence Index v5.0 includes: AA-Briefcase v1.1, AutomationBench-AA, Terminal-Bench 4.0, FrontierCode 2.0, AA-Omniscience v1.1, GDP.pdf, HLE-Diamond, CritPt v2 (Hard Subset), Terminal-Bench-Science 0.1. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Intelligence Index 与每项任务的成本

Artificial Analysis Intelligence Index · Weighted average cost (USD) per Artificial Analysis Intelligence Index task

Most attractive quadrant

Claude Haiku 5.5

Claude Opus 5.5

Pareto line

AA-Briefcase v1.1

AutomationBench-AA

单任务成本（美元，对数刻度）

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

成本

每项 Intelligence Index 任务的成本

Weighted average cost (USD) per Intelligence Index task · Segmented by token type · Lower is better

速度与延迟

输出速度

Output tokens per second · Higher is better

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API or the median across providers where a first-party API is not available.

能力得分

能力指数

衡量模型在特定能力与行业中的表现

Finance & Accounting Index

Incorporates 5 evaluations: AA-Briefcase v1.1, AA-Omniscience v1.1, HLE-Diamond, AutomationBench-AA, GDP.pdf · Higher is better

Strategy & Ops Index

Incorporates 4 evaluations: AutomationBench-AA, AA-Briefcase v1.1, AA-Omniscience v1.1, GDP.pdf · Higher is better

Legal Index

Incorporates 5 evaluations: AA-Omniscience v1.1, AA-Briefcase v1.1, HLE-Diamond, GDP.pdf, AutomationBench-AA · Higher is better

Healthcare & Medical Index

Incorporates 5 evaluations: AA-Omniscience v1.1, AA-Briefcase v1.1, MLCR-AA, AutomationBench-AA, HLE-Diamond · Higher is better

Engineering Index

Incorporates 7 evaluations: AA-Omniscience v1.1, CritPt v2 (Hard Subset), Terminal-Bench-Science 0.1, HLE-Diamond, Terminal-Bench 4.0, FrontierCode 2.0, AA-Briefcase v1.1 · Higher is better

Economics Index

Incorporates 4 evaluations: AA-Briefcase v1.1, AA-Omniscience v1.1, HLE-Diamond, GDP.pdf · Higher is better

基准测试

AA-PersonalityBenchNew

模型 Avs模型 B

AA-PersonalityBench：Expressiveness 指数，Claude Haiku 5.5 (max) vs Claude Opus 5.5 (max with fallback)

根据两者的 Elo 差距，每个模型的回答在各特质上被判定为更显著的概率 · 50% 表示势均力敌

Claude Haiku 5.5 (max)

Claude Opus 5.5 (max with fallback)

Metaphorical

21%

79%

Lexically Varied

21%

79%

Eloquent

24%

76%

Natural Sounding

59%

41%

Literate

51%

49%

Claude Haiku 5.5 vs. Claude Opus 5.5：发布比较

智能

速度

每任务成本

智能Updated

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index

Intelligence Index 与每项任务的成本

Cost per Intelligence Index Task

Artificial Analysis Intelligence Index

成本

每项 Intelligence Index 任务的成本

Cost per Intelligence Index Task

速度与延迟

输出速度

Output Speed

Model Performance Representation

能力得分

能力指数

基准测试

AA-PersonalityBenchNew

AA-PersonalityBench：Expressiveness 指数，Claude Haiku 5.5 (max) vs Claude Opus 5.5 (max with fallback)