跳轉到主要內容

Claude Haiku 5.5 (Max)

AnthropicClaudeProprietary

描述

Claude Haiku 5.5 is Anthropic's cheapest, fastest, and most capable small model to date, and the successor to Claude Haiku 4.5. It targets high-volume, cost-sensitive work (summaries, compaction, classification, database queries) and pairs with Opus 5.5 / Sonnet 5.5 as a coding subagent; it is especially strong for speed-sensitive tasks like live support and browser use. It is the first Haiku-class model with adjustable effort (low/medium/high/xhigh/max; Claude Platform default medium). It accepts text and image input (including file/PDF) with text output, supports multilingual and vision workloads, tool use, and adaptive thinking. Context window is 1M tokens with 128K max output; knowledge cutoff is June 2026. Claude Platform pricing is tiered by prompt length: prompts up to 100K tokens are $0.10/$0.50 per million input/output (cache read $0.01, 5m cache write $0.125); prompts over 100K are $0.50/$2.50 (cache read $0.05, 5m cache write $0.625). Catalog provider pricing stores the <=100K tier. Capability scores are self-reported from the Claude Haiku 5.5 System Card (Table 8.1.A and §8) and the launch announcement. Available on the Claude API as `claude-haiku-5-5`.

發布日期
2026-10-07
參數規模
—
上下文長度
1.0M
支援模態
image, pdf, text

能力雷達圖

44
general
55
coding
50
reasoning
51
science
30
agents
60
multimodal

排行榜排名

領域#排名分數來源
智慧體能力模型榜42
50.0
LS
程式碼能力榜15
93.0
AA
通用能力榜23
74.0
AA
數學推理39
86.0
LB
多模態榜83
50.0
LS
推理能力29
85.0
LB
科學能力49
76.0
AA

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

Agents

AA-Briefcase v1.11578.00 / 3000自報
Program Bench82.0%自報
OfficeQA73.5%自報
OfficeQA Pro60.3%自報

Biology

LatchBio SpatialBench Verified67.7%自報
Protocols Understanding (V2)64.1%自報
Protocols Troubleshooting58.0%自報
Medicinal Chemistry (ADME)58.0%自報
Protein Design — Library Ranking45.3%自報
Biomedical image analysis43.5%自報
Protein Design — Sequence Generation33.1%自報
Morphology-to-molecule matching25.4%自報

Code

BenchCAD (with Python tool)87.0%自報
BenchCAD67.0%自報
FrontierSWE V243.8%自報

General

GDPval-AA 2.11620.00 / 3000自報
Global-MMLU87.8%自報

Healthcare

HealthBench Professional (raw)71.0%自報
HealthBench Professional64.8%自報
HealthBench61.6%自報
PhysicianBench43.0%自報

Language

MILU87.6%自報

Multimodal

OSWorld 2.1 (partial)72.4%自報
OSWorld 2.1 (strict)37.1%自報

Reasoning

SWE-bench Multilingual83.7%自報
SWE-Bench ProPrinceton NLP (2024)64.8%自報
FrontierCode 1.1 (Extended)58.4%自報
Humanity's Last Exam (with tools)57.4%自報
FrontierCode 1.146.4%自報
Humanity's Last Exam (no tools, text-only)45.9%自報
Terminal-Bench 4.039.2%自報
SWE-Bench Multimodal30.7%自報
Terminal-Bench-Science 0.120.6%自報

Science

LatchBio SingleCellBench56.2%自報

Vision

Chartography86.2%自報
Chartography (no tools)46.4%自報

AA 評測指數

(Artificial Analysis)
Lcr(Artificial Analysis)
82.7
Scicode(UIUC + Argonne National Lab (2024))
55.0
Hle(Center for AI Safety + Scale AI (2025))
44.4
Intelligence Index(Artificial Analysis)
43.4
Terminalbench V4 0
32.8

LLM Stats 分類評分

(LLM Stats (zeroeval))
Productivity
100
Agents
100
Reasoning
100
General
100
Language
90
Multimodal
60
Science
60
Healthcare
60
Biology
60
Code
60
Vision
60
Math
50
Tool Calling
30

定價

輸入價格$0.1 / 1M tokens
輸出價格$0.5 / 1M tokens
混合價格(3:1)$0.2 / 1M tokens
快取讀取價格$0.01 / 1M tokens
快取寫入價格$0.125 / 1M tokens

速度

Tokens/秒227.9
首Token延遲292.43s
首回答延遲292.43s

供應商價格排行

供應商價格排行

1 個供應商

供應商輸入輸出
1Anthropic
$0
$0

比較該模型在不同 API 供應商之間的定價。

外部連結