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
6
AA 评分
53.9
官方来源
3 个补充入口
案例来源
6 条可核验
来自公开资料与模型数据库 + Artificial Analysis + 3 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
Kimi K2.6 当前关联 6 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
Kimi K2.6 的 6 条 A 类案例中,已有 6 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
3 个入口
6 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 | Kimi / Moonshot AI |
|---|---|
| 厂商页 | Kimi / Moonshot AI |
| 发布时间 | 2026-04-20 |
| 模型 ID | kimi-k2-6 |
| 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 | Kimi / Moonshot AI |
| AA 评分 | 53.9 |
| 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 | 6 |
谱系位置
它在厂商路线中的位置
此页将 Kimi K2.6 放入 Kimi / Moonshot AI 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。
2026-04-20
Kimi K2.6 资料状态
来自公开资料与模型数据库 + Artificial Analysis + 3 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
Kimi K2.6 当前标记为已有真实案例。
案例库
公开案例复核
已有 6 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。
2026-06-26T16:06:42.969375+00:00
Kimi K2.6 builds a full-stack FlowBoard project management app in the same PRD challenge as Claude Opus 4.6, GPT-5.4, and GLM 5.1
ZacharyZhang-NY / full_stack_web_development
2026-06-26T16:06:42.969341+00:00
Kimi K2.6 builds a SF6-style fighting game with 20 characters via OpenCode multi-agent workflow
matsumotokohei / game_development
2026-06-26T15:20:00Z
Talent Intelligence Agent — LinkedIn candidate discovery and scoring with Kimi K2.6
yashb98 / hr_tech
2026-06-26T15:20:00Z
Phantom Trader — Autonomous AI swing trading bot analyzing 1400+ stocks with Kimi K2.6
vinilpolepalli / fintech
2026-06-26T15:20:00Z
Le Analyst — CLI equity research assistant with WeChat mini-app using Kimi K2.6
Yogurto710 / fintech
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 6 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 6。
Representative Evidence
代表案例排序
当前第一代表案例是“Kimi K2.6 builds a full-stack FlowBoard project management app in the same PRD challenge as Claude Opus 4.6, GPT-5.4, and GLM 5.1”。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/kimi-k2-6-flowboard-showdown-prd/manifest.json
Coding / repo / agent
ZacharyZhang-NY 使用 Kimi K2.6 处理软件工程任务执行
ZacharyZhang-NY · Kimi K2.6
A+ 完整链路 · 代码仓库证据
ZacharyZhang-NY 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 2026-06-26T16:06:42.969375+00:00。 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 developer (ZacharyZhang-NY) ran the same detailed PRD (FlowBoard - a team project management board with drag-and-drop Kanban, Sprint planning, and Issue tracking) through 4 different AI coding models: GLM 5.1, Kimi K2…
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:A complete Next.js 16 application codebase in the Kimi-K2.6 directory including app routes, backend API, database schema, components, hooks, types, configuration files (next.config.ts, drizzle.config.ts, tsconfig.json),…
模型作用:Kimi K2.6 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:Kimi K2.6 generated the entire full-stack application from a single PRD prompt, including frontend components with IBM Carbon Design System, backend API routes, database schema with Drizzle ORM, authentication with Bett…
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:Repo has 2 stars; this is a model comparison project so Kimi K2.6 output is presented alongside other models; the PRD explicitly targets Kimi Code as the coding agent; Docker deployment suggests the generated code is fu…
原始记录:Kimi K2.6 builds a full-stack FlowBoard project management app in the same PRD challenge as Claude Opus 4.6, GPT-5.4, and GLM 5.1
matsumotokohei 使用 Kimi K2.6 处理可玩交互原型构建
matsumotokohei · Kimi K2.6
A+ 完整链路 · 代码仓库证据
matsumotokohei 公开的游戏与交互原型案例,来源为 公开代码库,复核于 2026-06-26T16:06:42.969341+00:00。 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:这是一个围绕可玩交互原型构建的真实任务,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 Japanese developer (matsumotokohei) used Kimi K2.6 as the primary coding model via OpenCode to build a Street Fighter 6-style 2D fighting game in Godot 3.6 GDScript. The project features 7 specialized AI sub-agents (m…
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:A complete playable Godot 3.6 fighting game project with 25+ GDScript files totaling ~350KB of game logic, including BattleManager (20KB), CPUController (36KB), CharacterData (35KB), Fighter (49KB), FighterVisual (56KB)…
模型作用:Kimi K2.6 在该案例中承担可玩交互原型构建相关的生成、分析、编排或实现角色。 原始资料写作:Kimi K2.6 served as the sole coding model powering all 7 AI agents in the OpenCode workflow: the manager agent handled requirements and task delegation, the developer agent implemented GDScript code, the reviewer agent …
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:Repo has 0 stars as of collection date; model is configured at OpenCode system level not in repo config files, but repo description explicitly states 'Model Kimi K2.6'; project appears to be a personal hobby project by …
原始记录:Kimi K2.6 builds a SF6-style fighting game with 20 characters via OpenCode multi-agent workflow
Fuhad-Lab 使用 Kimi K2.6 处理代码审查和测试生成
Fuhad-Lab · Kimi K2.6
A+ 完整链路 · 代码仓库证据
Fuhad-Lab 公开的代码审查与测试案例,来源为 公开代码库,复核于 2026-06-26T15:20:00Z。 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:Build a Render-hosted backend service for the Blue Horizon e-learning platform that uses Kimi K2.6 to generate and edit interactive HTML learning modules from natural language prompts, and Playwright to visually test an…
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:A deployed Render web service with 5 endpoints: /health (status check), /generate (create module from prompt), /edit (modify existing module), /test (visual testing with Playwright screenshots), /start and /action (brow…
模型作用:Kimi K2.6 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:Kimi K2.6 serves as the code generation engine, converting natural language teacher instructions into interactive HTML/CSS/JavaScript learning modules. Accessed via NVIDIA API. The model's coding capabilities enable it …
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:Low star count. Backend-only service with no public-facing demo URL. Uses NVIDIA API as inference provider. Blue Horizon platform deployment status unconfirmed.
原始记录:AI Module Tester — E-learning module generation and visual testing with Kimi K2.6
Yogurto710 使用 Kimi K2.6 处理软件工程任务执行
Yogurto710 · Kimi K2.6
A+ 完整链路 · 代码仓库证据
Yogurto710 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 2026-06-26T15:20:00Z。 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:Build a command-line research assistant for public equities that produces sourced, dated briefs on specific stock questions (analyst research TICKER QUESTION) and full 8-section initiation reports with peer comps, valua…
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:A Python CLI tool that produces equity research briefs (~$0.20-0.40/run, 2-3 min) and full initiation reports (~$0.60-0.80/run, 8-12 min). Output includes YAML frontmatter with ticker, date, language, model, and open qu…
模型作用:Kimi K2.6 在该案例中承担软件工程任务执行相关的生成、分析、编排或实现角色。 原始资料写作:Kimi K2.6 is the default LLM for all research and initiation report generation. It natively writes in both English and Chinese on the first pass, preserving financial notation ($, tickers, URLs) byte-for-byte in Chinese…
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:Low star count. WeChat mini-app companion mentioned but not confirmed deployed. Uses Tavily for web search augmentation. Research output is not financial advice.
原始记录:Le Analyst — CLI equity research assistant with WeChat mini-app using Kimi K2.6
vinilpolepalli 使用 Kimi K2.6 处理软件工程任务执行
vinilpolepalli · Kimi K2.6
A+ 完整链路 · 代码仓库证据
vinilpolepalli 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 2026-06-26T15:20:00Z。 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:Build an autonomous AI-powered swing trading system that runs nightly on GitHub Actions: analyze 1,400+ stocks using Kimi K2.6 via NVIDIA NIM API, score each stock 0-100 by buy conviction based on price momentum, sector…
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:A fully autonomous trading pipeline: 4 PM ET daily, Kimi K2.6 analyzes 1,437 stocks (~6 hours). 8:30 AM ET, a secondary system reads the report and places proportional buys on Robinhood. 3:30 PM ET checks +7%/-5% exit t…
模型作用:Kimi K2.6 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:Kimi K2.6 is the core analytical engine that evaluates each stock across 5 dimensions (price momentum, sector context, market conditions, technical setup, catalysts) and returns structured scores and signals. Accessed v…
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:Low star count. Automated trading involves financial risk. No confirmation of live trading performance or profitability. Uses NVIDIA NIM as inference provider, not direct Moonshot API.
原始记录:Phantom Trader — Autonomous AI swing trading bot analyzing 1400+ stocks with Kimi K2.6
yashb98 使用 Kimi K2.6 处理知识检索和问答
yashb98 · Kimi K2.6
A+ 完整链路 · 代码仓库证据
yashb98 公开的知识库与检索问答案例,来源为 公开代码库,复核于 2026-06-26T15:20:00Z。 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:Build an AI talent intelligence agent that takes a job spec, discovers LinkedIn candidates through the user's real Chrome session (via Kimi WebBridge), scrapes profiles, scores them on a 5-gate weighted rubric (Skills 3…
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:A complete Bun+TypeScript application with web dashboard, SQLite persistence, Notion API integration, and Telegram notifications. The agent autonomously runs 3-5 SearchWeb queries to collect 20-30 LinkedIn profiles, sco…
模型作用:Kimi K2.6 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:Kimi K2.6 serves as the core LLM powering the agent's autonomous planning, tool selection, candidate scoring logic, and personalized message drafting. It is accessed via @moonshot-ai/kimi-agent-sdk driving the local kim…
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:Repo has low star count (new project). Uses Kimi WebBridge which requires user's real Chrome session running locally. No public deployment URL confirmed. Model usage is via Kimi CLI OAuth, not direct API.
原始记录:Talent Intelligence Agent — LinkedIn candidate discovery and scoring with Kimi K2.6
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。