DATA CUT 2026-07-27 116 活跃模型 680 A 类案例

模型档案

Claude 3 Haiku

Claude 3 Haiku 已有可核验 A 类案例,可优先进入选型对比。

已有真实案例

档案摘要 / Decision snapshot

先判断能不能用于选型。

证据完整度 100%
公开状态 已有真实案例

可进入对比

案例覆盖 5 条 A 类

有真实任务证据

主要适用 作为完整模型索引和厂商路线追踪入口

来自模型卡

谨慎场景 缺少真实案例时不要包装成推荐

边界优先

30 秒结论

可进入选型对比,但仍以案例证据为准。

完整样板模型:已有官方来源、风险说明和可核验 A 类案例。

这页有来源、有风险说明、有公开可核验 A 类案例。

数据可信度 / 证据强度

已有真实案例

A 类案例

5

AA 评分

暂无数据

官方来源

3 个补充入口

案例来源

5 条可核验

来自厂商页/模型卡骨架 + 3 个厂商/官方/模型家族证据入口;暂无 Artificial Analysis 行。

Evidence Distribution

案例证据分布

Claude 3 Haiku 当前关联 5 条 A 类案例;这里按任务、来源和复核时间观察证据结构。

A CASES 5 5 selected

Evidence Snapshot

证据快照覆盖

Claude 3 Haiku 的 5 条 A 类案例中,已有 5 条完成本地证据快照;0 条仍在快照队列。

SNAPSHOT 100%
已快照 5

可用 manifest 复核

待快照 0

等待落盘

部分快照 0

需补抓

需关注 0

需人工复核

证据完整度

5 / 5

模型卡 OK

Feishu Bitable

官方/厂商来源 OK

3 个入口

A 类案例 OK

5 条

公开产物 OK

已绑定

风险备注 OK

已标注

能力边界

用标签表达信号,不用标签替代证据。

查看证据方法
Coding Agent

通用/待核验

长上下文

暂无数据

研究任务

待核验

开放生态

平台/待核验

生命周期风险

厂商Anthropic / Claude
厂商页Anthropic / Claude
发布时间2024-03-13
模型 IDclaude-3-haiku
上下文暂无数据
输出暂无数据
模态文本 / 多模态待核验
推理否或官方未披露
价格官方未披露 / 暂无数据
可用平台Anthropic / Claude
AA 评分暂无数据
基础模型发布时间Artificial Analysis 未披露
资料来源公开资料、厂商信息和案例库
A 类案例5

谱系位置

它在厂商路线中的位置

此页将 Claude 3 Haiku 放入 Anthropic / Claude 的当前 Atlas 路线中。若同厂前后代资料不足,先保留为路线追踪入口,不脑补谱系关系。

2024-03-13

Claude 3 Haiku 资料状态

来自厂商页/模型卡骨架 + 3 个厂商/官方/模型家族证据入口;暂无 Artificial Analysis 行。

公开档案

模型状态

Claude 3 Haiku 当前标记为已有真实案例。

案例库

公开案例复核

已有 5 条可核验 A 类案例,达到完整补齐线

案例时间线

最近进入档案的 A 类案例

用采集时间呈现证据进入 Atlas 的顺序,方便复核来源新鲜度。

适合场景

  • - 作为完整模型索引和厂商路线追踪入口
  • - 通用模型对比

不适合

  • - 缺少真实案例时不要包装成推荐
  • - 缺字段时不要脑补价格、性能或上下文

真实案例

可核验 A 类案例精选

当前展示 5 条代表案例;排序优先看模型卡精选、证据可信度、快照状态和公开仓库/技术任务信号。案例总数仍为 5。

Representative Evidence

代表案例排序

当前第一代表案例是“AWS Food Analyzer: Serverless Nutrition App Uses Claude 3 Haiku for Real-Time Ingredient Analysis and Recipe Generation”。它会优先出现在本页,是因为该案例同时进入模型卡精选、具备 100/100 的证据可信度,并带有 已快照证据。

打开第一代表案例
排序池 5 条

模型卡精选优先

证据可信度 100/100

A+ 完整链路

快照状态 已快照

archive/evidence/aws-food-analyzer-haiku/manifest.json

技术信号 命中

Coding / repo / agent

AWS Samples (AWS Hackathon France 2… 使用 Claude 3 Haiku 处理研究分析和报告生成

AWS Samples (AWS Hackathon France 2024 winner) · Claude 3 Haiku

A
厂商:Anthropic / Claude 模型:Claude 3 Haiku 来源平台:github 最后复核:2026-06-26T20:15:00Z 证据快照:已快照 · 2 / 2 个证据目标 归档说明:已完成快照:2 / 2 个证据目标,可用本地 manifest 复核证据指纹。
证据可信度 100/100

A+ 完整链路 · 代码仓库证据

原始证据1 个公开产物复核通过代码仓库证据

AWS Samples (AWS Hackathon France 2024 winner) 公开的研究与报告生成案例,来源为 公开代码库,复核于 2026-06-26T20:15:00Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。

任务:这是一个围绕研究分析和报告生成的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:Personalized nutritional analysis of grocery products via barcode scanning and food photo recognition; generate ingredient explanations, quantitative allergen detection, and personalized recipe suggestions based on user…

公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Production serverless app deployed on AWS (Lambda + DynamoDB + S3 + CloudFront + Bedrock); winner of AWS Hackathon France 2024; showcased at multiple AWS Summit events; processes product barcodes and food photos with pe…

模型作用:Claude 3 Haiku 在该案例中承担研究分析和报告生成相关的生成、分析、编排或实现角色。 原始资料写作:Claude 3 Haiku was explicitly chosen as 'a good ratio between latency and quality' for real-time food ingredient analysis, allergen detection, and recipe generation. Handles multi-shot prompt engineering for generating …

A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。

风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:AWS sample/demo project built for hackathon and summit demos, not a commercial product with paying customers. However, it is a real deployed application with production-grade serverless architecture and demonstrates Cla…

原始记录:AWS Food Analyzer: Serverless Nutrition App Uses Claude 3 Haiku for Real-Time Ingredient Analysis and Recipe Generation

已有真实案例 研究与报告生成公开代码库A 类可核验real_case auto_approved 进入模型卡精选

Lemony AI (lemony-ai) 使用 Claude 3 Haiku 处理智能体流程编排

Lemony AI (lemony-ai) · Claude 3 Haiku

A
厂商:Anthropic / Claude 模型:Claude 3 Haiku 来源平台:github 最后复核:2026-06-26T13:17:46Z 证据快照:已快照 · 2 / 2 个证据目标 归档说明:已完成快照:2 / 2 个证据目标,可用本地 manifest 复核证据指纹。
证据可信度 100/100

A+ 完整链路 · 代码仓库证据

原始证据1 个公开产物复核通过代码仓库证据

Lemony AI (lemony-ai) 公开的智能体工作流案例,来源为 公开代码库,复核于 2026-06-26T13:17:46Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。

任务:这是一个围绕智能体流程编排的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:CascadeFlow is an agent runtime intelligence layer (2,740 GitHub stars) that dynamically selects the optimal LLM for each query or tool call through speculative execution. It uses Claude 3 Haiku as the initial speculati…

公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Per-query optimal model routing decisions; 40-85% cost reduction while retaining 96% quality of flagship models; sub-5ms routing overhead in agent execution loops

模型作用:Claude 3 Haiku 在该案例中承担智能体流程编排相关的生成、分析、编排或实现角色。 原始资料写作:Claude 3 Haiku serves as the primary speculative execution model, handling the majority of queries at $0.15-0.30/1M tokens before potential escalation to more expensive models. This makes it the cost-efficiency backbone…

A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。

风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:CascadeFlow is a framework/tool that uses Haiku as part of its multi-model cascading strategy, not a standalone Haiku product deployment. The actual production impact depends on individual users' configurations and task…

原始记录:CascadeFlow: Claude 3 Haiku as Speculative Execution Model in Agent Cascading Runtime

已有真实案例 智能体工作流公开代码库A 类可核验real_case auto_approved 进入模型卡精选

vxcontrol 使用 Claude 3 Haiku 处理代码审查和测试生成

vxcontrol · Claude 3 Haiku

A
厂商:Anthropic / Claude 模型:Claude 3 Haiku 来源平台:github 最后复核:2026-06-26T13:17:46Z 证据快照:已快照 · 2 / 2 个证据目标 归档说明:已完成快照:2 / 2 个证据目标,可用本地 manifest 复核证据指纹。
证据可信度 100/100

A+ 完整链路 · 代码仓库证据

原始证据1 个公开产物复核通过代码仓库证据

vxcontrol 公开的代码审查与测试案例,来源为 公开代码库,复核于 2026-06-26T13:17:46Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。

任务:这是一个围绕代码审查和测试生成的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:PentAGI is a fully autonomous penetration testing platform (17,973 GitHub stars) that uses AI agents to perform complex security assessment tasks. It explicitly lists claude-3-haiku-20240307 as a supported LLM provider …

公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Comprehensive vulnerability reports with exploitation guides; automated penetration test execution across network reconnaissance, web application testing, and privilege escalation phases

模型作用:Claude 3 Haiku 在该案例中承担代码审查和测试生成相关的生成、分析、编排或实现角色。 原始资料写作:Claude 3 Haiku provides fast, cost-effective AI reasoning for real-time security analysis, automated task planning, and intelligent tool selection during penetration testing. Its speed (21K tokens/second) enables rapid …

A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。

风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:PentAGI supports 10+ LLM providers; Claude 3 Haiku is one option among many. Users can configure any supported provider. The pentagi-3-haiku-20240307 model is noted as being retired April 19, 2026, with migration to cla…

原始记录:PentAGI: Claude 3 Haiku for Automated Penetration Testing AI Agent Reasoning

已有真实案例 代码审查与测试公开代码库A 类可核验real_case auto_approved 进入模型卡精选

Omni AI (getomni-ai) 使用 Claude 3 Haiku 处理知识检索和问答

Omni AI (getomni-ai) · Claude 3 Haiku

A
厂商:Anthropic / Claude 模型:Claude 3 Haiku 来源平台:github 最后复核:2026-06-26T13:17:46Z 证据快照:已快照 · 2 / 2 个证据目标 归档说明:已完成快照:2 / 2 个证据目标,可用本地 manifest 复核证据指纹。
证据可信度 100/100

A+ 完整链路 · 代码仓库证据

原始证据1 个公开产物复核通过代码仓库证据

Omni AI (getomni-ai) 公开的知识库与检索问答案例,来源为 公开代码库,复核于 2026-06-26T13:17:46Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。

任务:这是一个围绕知识检索和问答的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:Zerox is a document OCR library (12,243 GitHub stars) that converts PDFs, DOCX files, and images into Markdown by passing pages as images to vision models. It explicitly lists Claude 3 Haiku (2024.03, 2024.10) as a supp…

公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Structured Markdown output from PDF/image documents; used for downstream LLM ingestion, data extraction pipelines, and document search systems

模型作用:Claude 3 Haiku 在该案例中承担知识检索和问答相关的生成、分析、编排或实现角色。 原始资料写作:Claude 3 Haiku provides fast, cost-effective vision-based OCR as one of Zerox's Bedrock-supported models. At $0.25/$1.25 per million tokens (input/output), it enables processing large document volumes at a fraction of t…

A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。

风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Zerox defaults to OpenAI GPT-4o for OCR; Claude 3 Haiku is one of several supported providers. Users must explicitly configure the Bedrock/Haiku provider. The product supports Anthropic direct API and AWS Bedrock, but t…

原始记录:Zerox OCR: Vision-based Document Extraction Using Claude 3 Haiku via AWS Bedrock

已有真实案例 知识库与检索问答公开代码库A 类可核验real_case auto_approved 进入模型卡精选

Clay 使用 Claude 3 Haiku 处理智能体流程编排

Clay · Claude 3 Haiku

A
厂商:Anthropic / Claude 模型:Claude 3 Haiku 来源平台:Anthropic customer story; product site 最后复核:2026-06-26 证据快照:已快照 · 2 / 2 个证据目标 归档说明:已完成快照:2 / 2 个证据目标,可用本地 manifest 复核证据指纹。
证据可信度 99/100

A+ 完整链路 · 官方/客户故事

原始证据1 个公开产物复核通过官方/客户故事

Clay 公开的智能体工作流案例,来源为 Anthropic customer story; product site,复核于 2026-06-26。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。

任务:这是一个围绕智能体流程编排的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:Clay used Claude 3 Haiku for lead identification, enrichment and personalized sales messaging.

公开产物:公开材料提供原始证据链接和可访问产物,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Anthropic customer story connects the Claude Haiku family to live Clay growth workflows.

模型作用:Claude 3 Haiku 在该案例中承担智能体流程编排相关的生成、分析、编排或实现角色。 原始资料写作:The story identifies Claude 3 Haiku as a top choice model for cost-effective outreach tasks.

A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。

风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Clay may route across models as Anthropic updates the Claude family; keep this as dated Claude 3 Haiku evidence.

原始记录:Clay sales enrichment and messaging

已有真实案例 智能体工作流Anthropic customer story; product siteA 类可核验real_case auto_approved 进入模型卡精选

数据缺口

只基于现有字段判断缺口;缺失项不会被猜测填充。

价格:待补公开资料 上下文:待补公开资料 输出:待补公开资料 官方/家族来源:已有 发布时间:已有 可用平台:已有 A 类案例:已有
  • - 补充方向:核验 价格 的官方/API/案例来源。
  • - 补充方向:核验 上下文 的官方/API/案例来源。
  • - 补充方向:核验 输出 的官方/API/案例来源。

下一步验证建议

  • - 继续保留 A 类案例的原始证据、产物页和版本快照。
  • - 复核官方来源与模型 ID、发布时间、平台可用性是否仍一致。
  • - 把 benchmark、教程、发布文保留为背景资料,不提升为真实案例。