MiMo-V2-Flash (Feb 2026)

herointene (Discord bot developer) 公开的翻译与本地化案例，来源为 公开代码库，复核于 2026-06-26T22:55:00Z。 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：这是一个围绕翻译和本地化处理的真实任务，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：Build a Discord bot that provides context-aware multilingual translation between Chinese, Japanese, and English, with automatic AI task recognition (summarization, email writing, etc.), using MiMo-V2-Flash as the sole 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-ready Discord bot deployed via Docker Compose with: (1) Flag emoji reaction triggers (🇨🇳/🇯🇵/🇬🇧) for forced target-language translation, (2) thread-based interaction to keep channels clean, (3) smart c…

模型作用：MiMo-V2-Flash (Feb 2026) 在该案例中承担翻译和本地化处理相关的生成、分析、编排或实现角色。 原始资料写作：MiMo-V2-Flash is the core engine powering all AI capabilities: multilingual translation, context understanding from recent message history, intent classification (translation vs. task fulfillment), and generation of tas…

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：No stars or external adoption metrics visible yet. The bot is functional but appears to be an individual developer project without documented production deployment scale.

原始记录：Discord AI Translator: Context-Aware Multilingual Bot Powered by MiMo-V2-Flash

blessblissmari (academic tool developer) 公开的代码审查与测试案例，来源为 公开代码库，复核于 2026-06-26T22:55:00Z。 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 5-pass cascade translation pipeline that converts English academic PDF/DOCX/PPTX documents into Russian DOCX/PPTX in MIET university template format, where MiMo-V2-Flash serves as the cost-efficient OCR comparis…

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 application (v1 live at blessblissmari.github.io/miet-translator-pro/) with a v2 cascade pipeline: Pass 1 (mimo-v2.5, global OCR layout map), Pass 2 (mimo-v2-omni, per-page OCR), Pass 3 (MiMo-V2-Flash, diff compar…

模型作用：MiMo-V2-Flash (Feb 2026) 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：MiMo-V2-Flash is used as the 'CHEAP' tier in a sophisticated multi-model pipeline, handling two critical roles: (1) Pass 3 'SVERKA' — diff-comparing OCR output against pdf.js text extraction to catch translation artifac…

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：Live demo at blessblissmari.github.io/miet-translator-pro/ returned empty response during verification (may be temporarily down or require JavaScript). GitHub repo README and source code structure are fully verifiable. …

原始记录：MIET Translator Pro v2: Academic Document Translation Pipeline Using MiMo-V2-Flash as QA Layer

serenemm (Roblox developer) 公开的游戏与交互原型案例，来源为 公开代码库，复核于 2026-06-26T22:55:00Z。 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：Deploy an LLM-powered NPC named 'Sam' inside a Roblox game that holds real-time natural language conversations with players via in-game chat bubbles, using MiMo-V2-Flash as the inference backend served through OpenRoute…

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 working Roblox game with an AI NPC ('Sam') that receives player chat messages through HttpService, sends them to a Railway-hosted backend, queries xiaomi/mimo-v2-flash:free via OpenRouter, and returns contextual conve…

模型作用：MiMo-V2-Flash (Feb 2026) 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：MiMo-V2-Flash provides the core conversational intelligence: it processes natural language player input and generates contextually appropriate NPC responses. The model was chosen for its free tier availability on OpenRo…

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：README architecture section mentions 'deepseek LLM model' in one place, likely a copy-paste error from a template; the explicit model declaration throughout the README and setup instructions consistently states xiaomi/m…

原始记录：Project Sam: AI-Powered Roblox NPC Conversations via MiMo-V2-Flash

AndreaPanzeriDev (game developer) 公开的游戏与交互原型案例，来源为 公开代码库，复核于 2026-06-26T22:55:00Z。 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：Use MiMo-V2-Flash to generate 100% of the source code for a fully playable 2D tower defense game, including 4 tower types, 3 enemy types, wave management, economy system, and Docker deployment configuration.

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 2D tower defense game with: 4 tower types (Basic, Sniper, Splash, Slow), 3 enemy types (Normal, Fast, Tank), increasing-difficulty wave system with UI, economy system (money from kills, tower costs/upgrades/s…

模型作用：MiMo-V2-Flash (Feb 2026) 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：MiMo-V2-Flash generated the entire codebase including game logic, rendering, UI, enemy AI, tower mechanics, wave spawning, economy balancing, and Docker configuration. The developer used a single prompt ('realize a 2d g…

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：Zero stars; appears to be a personal demo project. The '100% coded' claim is asserted by the developer but the repo contains only a Dockerfile and game code without explicit generation logs.

原始记录：Tower Defense Game 100% AI-Coded Using MiMo-V2-Flash

Alibaba Group / OpenCodeReview 公开的代码审查与测试案例，来源为 公开代码库，复核于 2026-06-26T14:10:34Z。 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：Alibaba's OpenCodeReview (9,380 stars), an AI-powered code review CLI tool battle-tested at Alibaba's scale serving tens of thousands of developers, included MiMo-V2-Flash as a supported model in its built-in MiMo provi…

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：MiMo-V2-Flash was listed as a supported model in OpenCodeReview's built-in provider registry alongside other MiMo models (mimo-v2-pro, mimo-v2-omni). The model was removed in PR #192 (merged 2026-06-23) after Xiaomi ann…

模型作用：MiMo-V2-Flash (Feb 2026) 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：MiMo-V2-Flash served as one of the recommended models for AI-powered code review in Alibaba's OpenCodeReview tool. Its reasoning and coding capabilities made it suitable for the tool's hybrid architecture combining dete…

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：MiMo-V2-Flash deprecated from Xiaomi platform on 2026-06-30; removed from OpenCodeReview provider registry. Historical use case.

原始记录：Alibaba OpenCodeReview: MiMo-V2-Flash as Supported LLM for AI Code Review

theopenco (LLMGateway project) 公开的真实任务执行案例，来源为 公开代码库，复核于 2026-06-26T14:10:34Z。 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：这是一个围绕真实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：LLMGateway (1,338 stars), a unified API interface for routing, managing, and analyzing LLM requests across multiple providers, had MiMo-V2-Flash configured as an active model route through the Xiaomi MiMo Open Platform …

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：MiMo-V2-Flash was an active routed model in LLMGateway's provider registry. When Xiaomi announced the model would be retired on 2026-06-18 (with silent auto-forwarding to MiMo-V2.5), LLMGateway proactively deactivated t…

模型作用：MiMo-V2-Flash (Feb 2026) 在该案例中承担真实任务执行相关的生成、分析、编排或实现角色。 原始资料写作：MiMo-V2-Flash was integrated as a production-ready model option in LLMGateway's multi-provider routing system, demonstrating its compatibility with standard OpenAI-compatible API interfaces and its viability as a routin…

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：MiMo-V2-Flash retired from Xiaomi MiMo Open Platform on 2026-06-18; model route deactivated. Historical use case.

原始记录：LLMGateway: MiMo-V2-Flash as Routed Model in Multi-Provider LLM Gateway

aaronjmars (MiroShark project) 公开的文档理解与处理案例，来源为 公开代码库，复核于 2026-06-26T14:10:34Z。 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：MiroShark (1,337 stars), a Universal Swarm Intelligence Engine for financial forecasting and future prediction, used xiaomi/mimo-v2-flash as its default LLM model via OpenRouter for persona generation, simulation config…

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 deployment as default model across multiple simulation slots (Default + Wonderwall). Each simulation run consumed 7M+ tokens across 850+ agent-action calls. The project documented MiMo-V2-Flash pricing at $0.…

模型作用：MiMo-V2-Flash (Feb 2026) 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：MiMo-V2-Flash's combination of strong reasoning capability and competitive pricing ($0.10/M input) made it viable as the default model for high-volume swarm intelligence simulations. The model's persona generation quali…

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：MiMo-V2-Flash was delisted from OpenRouter on 2026-06-22; project switched to MiMo-V2.5. Historical use case.

原始记录：MiroShark: MiMo-V2-Flash as Default Model for Swarm Intelligence Simulations

LMSYS / SGLang Project 公开的多模态生成与理解案例，来源为 公开代码库，复核于 2026-06-26T14:10:34Z。 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：这是一个围绕多模态内容处理的真实任务，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：SGLang team (contributor acelyc111) integrated MiMo-V2-Flash as a first-class supported model in the SGLang LLM serving framework, implementing optimized Sliding Window Attention (SWA) execution, multi-layer MTP (Multi-…

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：Merged PR #15207 with full day-0 support. Achieved 150 TPS per request decoding throughput even under 64K input tokens with batch size 16 per DP rank. Published detailed benchmarking results for prefill and decode phase…

模型作用：MiMo-V2-Flash (Feb 2026) 在该案例中承担多模态内容处理相关的生成、分析、编排或实现角色。 原始资料写作：MiMo-V2-Flash's hybrid SWA+GA attention architecture and 3-layer MTP design enabled SGLang to demonstrate that inference-centric model design can achieve balanced throughput and latency. The model's architecture was spe…

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：MiMo-V2-Flash was retired from Xiaomi MiMo Open Platform on 2026-06-18; model may no longer be available via official API. Historical use case.

原始记录：SGLang Day-0 Support for MiMo-V2-Flash with Optimized Inference