可进入对比
档案摘要 / Decision snapshot
先判断能不能用于选型。
有真实任务证据
来自模型卡
边界优先
30 秒结论
可进入选型对比,但仍以案例证据为准。
来自公开模型资料库,并已补厂商/官方/模型家族证据;当前站内暂无可核验 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
强
A 类案例
5
AA 评分
32.51
官方来源
2 个补充入口
案例来源
5 条可核验
来自公开资料与模型数据库 + Artificial Analysis + 2 个厂商/官方/模型家族证据入口;AA 不是唯一事实来源。
Evidence Distribution
案例证据分布
GLM-4.6 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。
Evidence Snapshot
证据快照覆盖
GLM-4.6 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。
可用 manifest 复核
等待落盘
需补抓
需人工复核
证据完整度
5 / 5
Feishu Bitable
2 个入口
5 条
已绑定
已标注
能力边界
用标签表达信号,不用标签替代证据。
通用/待核验
暂无数据
待核验
平台/待核验
中
| 厂商 | Z AI / GLM |
|---|---|
| 厂商页 | Z AI / GLM |
| 发布时间 | 2025-09-30 |
| 模型 ID | glm-4-6 |
| 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 | 是 / 推理模型或 thinking 模式 |
| 价格 | 官方未披露 / 暂无数据 |
| 可用平台 | Z AI / GLM |
| AA 评分 | 32.51 |
| 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 GLM-4.6 放入 Z AI / GLM 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。
2025-09-30
GLM-4.6 资料状态
来自公开资料与模型数据库 + Artificial Analysis + 2 个厂商/官方/模型家族证据入口;AA 不是唯一事实来源。
公开档案
模型状态
GLM-4.6 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。
2026-06-27T07:00:02Z
Xiexie: senior scam-shield macOS companion built for GOSIM Agentic Hackathon
Edouard Foussier / Xiexie / assistive agent / scam detection
2026-06-27T07:00:02Z
GLM MCP Server for delegating Claude Code tasks to GLM-4.6
yasinaky / developer tooling / coding agent integration
2026-06-27T07:00:02Z
CCGLM MCP Server routing Claude Code prompts to Z.AI GLM-4.6
CyPack / developer tooling / MCP routing
2026-06-27T07:00:02Z
Z.ai Coding Agent CLI using GLM-4.6 for large-codebase coding tasks
lokafinnsw / developer tooling / coding assistant
2026-06-27T07:00:02Z
GoogleReviewHelper web app coded with assistance from GLM-4.6
Auxiar / software development / web app generation
适合场景
- - 作为完整模型索引和厂商路线追踪入口
- - 推理/agentic workflow 候选
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“GoogleReviewHelper web app coded with assistance from GLM-4.6”。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。
模型卡精选优先
A+ 完整链路
archive/evidence/glm-4-6-auxiar-google-review-helper/manifest.json
Coding / repo / agent
Auxiar 使用 GLM-4.6 处理软件工程任务执行
Auxiar · GLM-4.6
A+ 完整链路 · 代码仓库证据
Auxiar 公开的代码代理与软件工程案例,来源为 公开代码库,复核于 2026-06-27T07:00:02Z。 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:这是一个围绕软件工程任务执行的真实任务,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:A web page that calls Google's Places API to retrieve a direct 'Leave a Review' Google link from a supplied Places API key.
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:Public GitHub repository and hosted GitHub Pages app for generating Google review links.
模型作用:GLM-4.6 在该案例中承担软件工程任务执行相关的生成、分析、编排或实现角色。 原始资料写作:The GitHub repository description states the webpage was quickly coded with help from GLM-4.6 and Claude, explicitly recommending GLM-4.6.
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:Evidence is from repository metadata rather than README text, and the author credits Claude alongside GLM-4.6; treat as co-assisted coding evidence.
原始记录:GoogleReviewHelper web app coded with assistance from GLM-4.6
CyPack 使用 GLM-4.6 处理软件工程任务执行
CyPack · GLM-4.6
A+ 完整链路 · 代码仓库证据
CyPack 公开的代码代理与软件工程案例,来源为 公开代码库,复核于 2026-06-27T07:00:02Z。 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:这是一个围绕软件工程任务执行的真实任务,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:A Spanish-language MCP server that exposes a glm_route tool so Claude Code can invoke another Claude CLI instance configured against the Z.AI Anthropic-compatible endpoint.
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:Public repository with MCP server code, architecture diagram text, environment configuration, and installation steps for routing prompts to GLM-4.6.
模型作用:GLM-4.6 在该案例中承担软件工程任务执行相关的生成、分析、编排或实现角色。 原始资料写作:The README explicitly lists the architecture path ending in 'Modelo GLM-4.6' and states glm_route routes prompts to GLM-4.6 for code generation, deep analysis, and general queries.
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:Runtime depends on the user's Z.AI API credentials; the public artifact demonstrates the integration and intended GLM-4.6 usage, not a hosted service.
原始记录:CCGLM MCP Server routing Claude Code prompts to Z.AI GLM-4.6
Edouard Foussier / Xiexie 使用 GLM-4.6 处理软件工程任务执行
Edouard Foussier / Xiexie · GLM-4.6
A+ 完整链路 · 代码仓库证据
Edouard Foussier / Xiexie 公开的代码代理与软件工程案例,来源为 公开代码库,复核于 2026-06-27T07:00:02Z。 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:这是一个围绕软件工程任务执行的真实任务,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:A voice-first macOS companion for seniors that reads the screen, answers spoken questions, points to UI elements, and helps identify scam emails.
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:Public hackathon-winning repository, landing page, and 3-minute demo showing Xiexie analyzing suspicious emails and guiding the user on screen.
模型作用:GLM-4.6 在该案例中承担软件工程任务执行相关的生成、分析、编排或实现角色。 原始资料写作:The README explicitly states the project was built on Z.AI GLM-4.6 plus GLM-4.5V; GLM-4.6 is used as part of the agent stack for long-horizon assistant reasoning and interaction, while the visual email-reading flow uses…
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:GLM-4.6 is one of multiple models/services in the stack; visual screenshot analysis is attributed to GLM-4.5V, so cite this case as a multi-model agent application rather than GLM-4.6-only vision evidence.
原始记录:Xiexie: senior scam-shield macOS companion built for GOSIM Agentic Hackathon
lokafinnsw 使用 GLM-4.6 处理软件工程任务执行
lokafinnsw · GLM-4.6
A+ 完整链路 · 代码仓库证据
lokafinnsw 公开的代码代理与软件工程案例,来源为 公开代码库,复核于 2026-06-27T07:00:02Z。 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:这是一个围绕软件工程任务执行的真实任务,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:A Rust command-line coding assistant for code explanation, documentation, refactoring suggestions, bug detection, and interactive or single-query coding help.
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:Public CLI repository with installation instructions, model selection, API-key setup, and documented commands for using the coding assistant.
模型作用:GLM-4.6 在该案例中承担软件工程任务执行相关的生成、分析、编排或实现角色。 原始资料写作:The README describes the tool as an AI-powered coding assistant using Z.ai's GLM-4.6 model and lists GLM-4.6 with a 200K context window for complex coding 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:Repository appears to be an independent implementation and not an official Z.AI repo despite README clone text referencing z-ai/zai-coding-agent; evidence still clearly binds the artifact to GLM-4.6 usage.
原始记录:Z.ai Coding Agent CLI using GLM-4.6 for large-codebase coding tasks
yasinaky 使用 GLM-4.6 处理软件工程任务执行
yasinaky · GLM-4.6
A+ 完整链路 · 代码仓库证据
yasinaky 公开的代码代理与软件工程案例,来源为 公开代码库,复核于 2026-06-27T07:00:02Z。 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:这是一个围绕软件工程任务执行的真实任务,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:An MCP server that lets Claude Code call Z.ai GLM models for subtasks, agent work, parallel processing, cross-validation, debugging, and cost optimization.
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:Public Python MCP server repository with setup instructions, smoke tests, API-key configuration, and documented use through Claude Code MCP.
模型作用:GLM-4.6 在该案例中承担软件工程任务执行相关的生成、分析、编排或实现角色。 原始资料写作:The README names GLM-4.6 as the flagship model for reasoning, coding, and agentic tasks and describes offloading subtasks to GLM-4.6 from Claude Code.
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:The artifact is an integration tool rather than an end-user product; evidence is strong for real runtime use of GLM-4.6 in coding-agent workflows.
原始记录:GLM MCP Server for delegating Claude Code tasks to GLM-4.6
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。
- - 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。
- - 补官方发布/API/System Card 来源;若官方未披露,继续标注“官方未披露”。
- - 把 benchmark、教程、发布文保留为背景资料,不提升为真实案例。