开始使用
API 接入
Haiku 5.5 API 迁移说明
发布页列出的模型 ID 为 claude-haiku-5-5；使用条件和计费以平台为准。
Aixy API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
Merge Gateway API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
NEAR AI Cloud API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
Eden AI API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
LLM Gateway API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
CrossModel API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
Perplexity Agent API文档
目录部署标识：anthropic/claude-haiku-5-5。具体区域、权限与价格以服务商为准。
Anthropic API文档
目录部署标识：claude-haiku-5-5。具体区域、权限与价格以服务商为准。
相关动态
订阅更新公开成绩与计算依据
综合参考分根据已公开的评测资料整理。查看发布日期依据
综合参考分 59.1 · 17项基准
推理求解 63.1
FrontierMath · 0.4634146341463415 比例
Epoch AI · 独立评测 · 2
claude-haiku-5-5_max · epoch-frontiermath-reviewed-method-20261029-task-version-2.1.0 · 2026-10-07公开
Terminal-Bench-Science · 10.00%
Vals AI · 独立评测 · 0.1.0
Claude Haiku 5.5 / Mini-SWE-agent · Current displayed evaluator row. Mini-SWE-agent. Scores are not pooled across harnesses; source methodology and repeat counts retained.
HLE · 0.296 比例
Artificial Analysis · 独立评测 · Diamond
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; hle-diamond
CritPt · 0.435 比例
Artificial Analysis · 独立评测 · 2 Hard
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; critpt-v2-hard
工具与任务执行 69.1
GDPval-AA · 1630.83 分
Artificial Analysis · 独立评测 · 2.1
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; gdpval-aa-normalized
Terminal-Bench · 0.328282828282828 比例
Artificial Analysis · 独立评测 · 4.0
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; terminalbench-4-0
AutomationBench · 0.35411015132783624 比例
Artificial Analysis · 独立评测 · current published task suite
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; automationbench-aa
AA-Briefcase · 1574.12 分
Artificial Analysis · 独立评测 · 1.1
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; aa-briefcase
OSWorld · 72.4% Offline subset
www.anthropic.com · 官方发布比较 · 2.1
Haiku 5.5 · Published page official-cdb1d51fae04; row Computer use OSWorld 2.1; preserve page footnotes and task-specific settings · 2026-10-07公开
AA-AnalystAgent · 0.5375 比例
Artificial Analysis · 独立评测 · current AA pass^5
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; aa-analyst-agent
编程 69.8
FrontierCode · 0.349325505050505 比例
Artificial Analysis · 独立评测 · 2.0
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; frontiercode-2-0
Vals Code Migration · 52.36%
Vals AI · 独立评测 · 1.1
Claude Haiku 5.5 / Published Code Migration 1.1 sandbox · Current displayed evaluator row. Published Code Migration 1.1 sandbox. Scores are not pooled across harnesses; source methodology and repeat counts retained.
语言与信息处理 35.2
GDP.pdf · 0.208 比例
Artificial Analysis · 独立评测 · current published task suite
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; gdp-pdf
SimpleQA Verified · 0.238 比例
Epoch AI · 独立评测 · run 1.2.0
claude-haiku-5-5_max · epoch-simpleqa-reviewed-method-20261029-task-version-1.2.0 · 2026-10-09公开
MultiChallenge · 67.67 %
Scale AI · 独立评测 · current published task suite
Claude Haiku 5.5 · scale-multichallenge-reviewed-method-20261029
AA-Omniscience · 0.3466666666666667 比例
Artificial Analysis · 独立评测 · 1.1
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; omniscience-accuracy
AA-Omniscience · 0.6576530612244897 比例
Artificial Analysis · 独立评测 · 1.1
Claude Haiku 5.5 (Max) · Artificial Analysis current published protocol; omniscience-non-hallucination
多模态理解 50.0
Chartography · 46.4% no tools
www.anthropic.com · 官方发布比较 · current published task suite
Haiku 5.5 · Published page official-cdb1d51fae04; row Visual reasoning Chartography; preserve page footnotes and task-specific settings · 2026-10-07公开