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
49.62
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
MiniMax-M2.7 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
MiniMax-M2.7 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 | MiniMax |
|---|---|
| 厂商页 | MiniMax |
| 发布时间 | 2026-03-18 |
| 模型 ID | minimax-m2-7 |
| 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 | MiniMax |
| AA 评分 | 49.62 |
| 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 MiniMax-M2.7 放入 MiniMax 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。
2026-03-18
MiniMax-M2.7 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
MiniMax-M2.7 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。
2026-06-26T16:00:00Z
MoonBags: MiniMax M2.7 as LLM Exit Advisor for Solana Meme-Token Auto-Trading
fciaf420 / financial-trading
2026-06-26T16:00:00Z
Agent-Pilot: 6-Agent Multi-Agent Pipeline for Feishu IM-to-Presentation Automation Using MiniMax M2.7
bcefghj / multi-agent-automation
2026-06-26T16:00:00Z
ModelDocUpgrade: Word Document Rule Compliance Checker Using MiniMax-M2.7-highspeed
xiaoyza / document-compliance
2026-06-26T16:00:00Z
cnsoftbei: Multi-Agent AI Personalized Learning Platform for China Software Cup Using MiniMax-M2.7
DanoAndHolidays / education-ai
2026-06-26T16:00:00Z
qq-codememory-bot: QQ Group AI Bot with Local Code Search and Long-Term Memory Powered by MiniMax M2
EricSun0218 / developer-tools
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“Agent-Pilot: 6-Agent Multi-Agent Pipeline for Feishu IM-to-Presentation Automation Using MiniMax M2.7”。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/m27-agent-pilot-feishu-multiagent/manifest.json
Coding / repo / agent
bcefghj 使用 MiniMax M2.7 处理软件工程任务执行
bcefghj · MiniMax M2.7
A+ 完整链路 · 代码仓库证据
bcefghj 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 2026-06-26T16:00:00Z。 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:Agent-Pilot is a multi-agent collaboration system built for the Feishu AI Campus Challenge. It uses 6 specialized AI Agents (Intent, Planner, Research, Writer, Review, Builder) with MiniMax M2.7 for tool calling, enabli…
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 operational multi-agent pipeline with live demo (http://8.136.98.175), 50-page technical whitepaper, MCP Server for Cursor/Claude Desktop integration, and Flutter+WebSocket real-time dashboard. 14/14 PRD tests p…
模型作用:MiniMax M2.7 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:MiniMax M2.7 provides the tool-calling backbone for the multi-agent system — each of the 6 agents uses M2.7 for reasoning and tool invocation. The model's coding and agent capabilities enable structured task decompositi…
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:Built for a specific competition (Feishu AI Campus Challenge). Live demo server at 8.136.98.175 may not be permanently available. 9 GitHub stars as of June 2026.
原始记录:Agent-Pilot: 6-Agent Multi-Agent Pipeline for Feishu IM-to-Presentation Automation Using MiniMax M2.7
DanoAndHolidays 使用 MiniMax M2.7 处理软件工程任务执行
DanoAndHolidays · MiniMax M2.7
A+ 完整链路 · 代码仓库证据
DanoAndHolidays 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 2026-06-26T16:00:00Z。 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:A multi-agent personalized learning platform built for the 15th China Software Cup (A3 track). The system uses multiple AI agents (profile construction, resource generation, path planning, intelligent tutoring, learning…
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 AI-driven personalized learning platform with 5 collaborative agents covering the full learning lifecycle: student profiling, resource generation, learning path planning, real-time tutoring, and learning asse…
模型作用:MiniMax M2.7 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:MiniMax-M2.7 serves as the foundational LLM powering all 5 learning agents. The model's coding and reasoning capabilities enable: (1) student profile construction from interaction data, (2) educational resource generati…
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:Competition submission for China Software Cup — may not be a production-deployed system. Uses Anthropic-compatible API format to call MiniMax. 2 GitHub stars.
原始记录:cnsoftbei: Multi-Agent AI Personalized Learning Platform for China Software Cup Using MiniMax-M2.7
xiaoyza 使用 MiniMax M2.7 处理文档理解和结构化处理
xiaoyza · MiniMax M2.7
A+ 完整链路 · 代码仓库证据
xiaoyza 公开的文档理解与处理案例,来源为 公开代码库,复核于 2026-06-26T16:00:00Z。 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:这是一个围绕文档理解和结构化处理的真实任务,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:ModelDocUpgrade is a bidding document compliance review tool that uses MiniMax-M2.7-highspeed to automatically extract rules from a rule document (e.g., bidding requirements) and check them against a response document (…
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 web-based document review tool (Node.js server + browser UI) that produces per-rule compliance assessments with four evidence states (met/partial/missing/risk) and an overall verdict (PASS/RISK/INCOMPLETE). Extracts r…
模型作用:MiniMax M2.7 在该案例中承担文档理解和结构化处理相关的生成、分析、编排或实现角色。 原始资料写作:MiniMax-M2.7-highspeed is the core reasoning engine that reads structured document content, understands complex compliance rules from bidding documents, and evaluates whether response documents satisfy each rule. The mo…
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:Uses MiniMax-M2.7-highspeed variant specifically. Built for a niche but practical document compliance use case. Small repo, no stars recorded.
原始记录:ModelDocUpgrade: Word Document Rule Compliance Checker Using MiniMax-M2.7-highspeed
fciaf420 使用 MiniMax M2.7 处理研究分析和报告生成
fciaf420 · MiniMax M2.7
A+ 完整链路 · 代码仓库证据
fciaf420 公开的研究与报告生成案例,来源为 公开代码库,复核于 2026-06-26T16:00:00Z。 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:MoonBags is a Solana meme-token auto-trading bot that consumes real-time discovery streams from OKX smart-money signals and GMGN curated trenches, executes swaps via Jupiter Ultra, and uses MiniMax M2.7 as an optional L…
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 production Solana meme-token trading bot with Telegram control interface and web dashboard, where MiniMax M2.7 provides real-time, data-driven sell decisions based on multi-signal on-chain analysis. The bot operates i…
模型作用:MiniMax M2.7 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:MiniMax M2.7 serves as the cognitive exit advisor, analyzing complex multi-dimensional on-chain data (smart money flow, developer token holdings, holder profit/loss, price klines) every 30 seconds to make time-sensitive…
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:M2.7 is an optional component; the bot also works with configurable trail/stop exits without LLM. Financial trading use case — disclaimer in repo states not financial advice. 34 GitHub stars as of June 2026.
原始记录:MoonBags: MiniMax M2.7 as LLM Exit Advisor for Solana Meme-Token Auto-Trading
EricSun0218 使用 MiniMax M2.7 处理软件工程任务执行
EricSun0218 · MiniMax M2.7
A+ 完整链路 · 代码仓库证据
EricSun0218 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 2026-06-26T16:00:00Z。 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:A QQ group AI assistant bot that uses MiniMax M2 (M2.7 family) as its brain, combined with local code repository search (via ripgrep) and SQLite/FTS5 long-term memory. When mentioned in a QQ group, the bot answers code …
公开产物:A production QQ group bot with: (1) agentic code search using ripgrep with sandbox-locked repo access, (2) automatic memory extraction from conversations with deduplication, (3) manual memory management (remember/forget/list commands), (4) FTS5 Chinese bigram fuzzy recall for memory search, (5) multi-key rotation for rate limit resilience. Supports commands like '@bot 鉴权在哪实现的?' for code search and '@bot 记住:部署统一用 fly deploy' for memory storage.
模型作用:MiniMax M2.7 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:MiniMax M2 (M2.7 family) acts as the reasoning brain with tool-calling capabilities: the model iteratively invokes rg_search, read_file, and list_dir tools to explore local codebases, synthesizes findings with long-term…
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:Uses NapCatQQ unofficial protocol bridge which violates Tencent ToS (repo warns to use alt account only). README references 'MiniMax M2' family rather than explicitly M2.7, but M2.7 is the latest model in the M2 family …
原始记录:qq-codememory-bot: QQ Group AI Bot with Local Code Search and Long-Term Memory Powered by MiniMax M2
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。