--- 格式版本: 2 标题: "Compatible MiniMax Models" 原文链接: "https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm" 发布日期: "2026-09-04" 发布时间校准状态: "found" 发布时间需复核: "否" 发布时间来源: "rule:local:strict_original_body" 发布时间证据: "Updated 2026-09-04" 发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body" 发布时间校准置信度: "high" 发布时间候选数量: 8 发布时间严格候选数量: 1 发布时间原页读取状态: "source template page reused from URL open" 发布时间未找到原因: "" 发布时间校准时间: "2026-09-05T23:12:26+08:00" 发布时间仲裁状态: "skipped" 发布时间仲裁尝试次数: 0 发布时间仲裁耗时毫秒: 0 发现时间: "2026-09-05T22:42:31+08:00" 入库时间: "2026-09-05T15:12:33.959Z" 来源平台: "固定入口" 搜索渠道: "fixed_url" 搜索词: "https://docs.oracle.com/en-us/iaas/releasenotes" 匹配关键词: - "AI" - "performance" 相关厂家: - "Oracle" 相关专家: [] 内容类型: "网页" 抓取工具: "Jina Reader" 清洗工具: "Jina Reader Markdown + Defuddle/Readability 正文提取" 原始附件: [] AI优质: "否" AI打分: 39 AI分档: "非优质" AI质检状态: "不通过" AI打分理由: "正文主线为Oracle官方文档，说明OCI Generative AI服务可导入MiniMax M3/M2模型并创建端点，属于模型兼容性和使用指南，非超节点/机架级AI基础设施。来源权威性高，但主题高度不相关；虽有模型参数、硬件shape等细节，但与机架级架构、互连、供电、液冷无关；无量产、客户、部署信号。历史知识库无同主题竞品，但本文无任何rack-scale新增事实。命中硬否决“教程与运维选型”：How to/选型文档，无新机架级硬件/协议标准，最高54分。" AI质检模型: "zj-deepseek-v4-flash" AI质检时间: "2026-09-05T23:12:49+08:00" AI主题相关性: 3 AI来源权威性: 13 AI新颖性: 7 AI技术细节: 6 AI商业部署信号: 2 AI完整性: 8 AI评分提示词版本: "v17-精简生产版" AI评分提示词SHA256: "48fb9777f386026761b4873eaff30807694fb11e9b352d7c69bf2dfde750cc7d" AI评分知识库版本: "knowledge_base_v1-20260819+runtime.93" AI评分知识库SHA256: "53d848ab04cf9f1c85be6ad1862e1223f16aca935d4b37b247fb402df13f2909" AI评分知识库检索词: "[\"Oracle\",\"https://docs.oracle.com/en-us/iaas/releasenotes\",\"RAS\",\"Meta\",\"URL\",\"UK\",\"EU\",\"OCI\",\"OHPPanel1nav0625\",\"M3\",\"minimax-m3\",\"M2\"]" AI评分知识库命中: "[{\"id\":\"july-correct-0001\",\"title\":\"全球首颗2nm GPU来了！苏姿丰甩出“最强AI机架”，CPU性能干翻英伟达 - 智东西\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"Oracle\",\"RAS\",\"Meta\",\"M3\",\"M2\"],\"rank\":-13.32136322214743},{\"id\":\"runtime-dccd09f98e2ca028070793f0\",\"title\":\"https://blogs.oracle.com/cloud-infrastructure/oci-achieves-nvidia-exemplar-cloud\",\"sourceType\":\"ai_excellent_article\",\"time\":\"\",\"matchedTerms\":[\"Oracle\",\"RAS\",\"OCI\"],\"rank\":-12.841073491450363},{\"id\":\"runtime-66d1ddc58948150bffee973a\",\"title\":\"Delivering Vera: NVIDIA’s First CPU Built for Agents Is Shipping Now\",\"sourceType\":\"ai_excellent_article\",\"time\":\"\",\"matchedTerms\":[\"Oracle\",\"RAS\",\"EU\",\"OCI\"],\"rank\":-11.327070846944585},{\"id\":\"july-correct-0015\",\"title\":\"AMD to join the optical interconnect party with 2027 Instinct GPUs\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"Meta\",\"URL\",\"OCI\"],\"rank\":-10.855958741112204},{\"id\":\"runtime-0210d17c29c694da454d0e78\",\"title\":\"HPE caps stellar Q3 with Oracle ‘gigawatt scale’ networking deal\",\"sourceType\":\"ai_excellent_article\",\"time\":\"\",\"matchedTerms\":[\"Oracle\",\"RAS\",\"OCI\"],\"rank\":-10.054298345295688}]" 采集批次: "2026年9月5日22点15分23秒" 采集批次ID: "20260905-221523-192bc80f" 去重键: "https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm" --- Title: Compatible MiniMax Models URL Source: https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm Markdown Content: * [Oracle Cloud Infrastructure Cloud Adoption Framework](https://docs.oracle.com/en-us/iaas/Content/cloud-adoption-framework/home.htm) * [Getting Started](https://docs.oracle.com/en-us/iaas/Content/GSG/Concepts/baremetalintro.htm) * [Oracle Cloud's Free Tier](https://docs.oracle.com/en-us/iaas/Content/FreeTier/freetier.htm) * [Oracle Multicloud](https://docs.oracle.com/en-us/iaas/Content/multicloud/Oraclemulticloud.htm) * [Oracle Dedicated Cloud](https://docs.oracle.com/en-us/iaas/Content/dedicated/home.htm) * [Oracle Cloud Infrastructure Government Cloud](https://docs.oracle.com/en-us/iaas/Content/gov-cloud/govlanding.htm) * [Oracle UK Sovereign Cloud](https://docs.oracle.com/en-us/iaas/Content/uk-sovereign-cloud/home.htm) * [Oracle EU Sovereign Cloud](https://docs.oracle.com/en-us/iaas/Content/sovereign-cloud/eu-sovereign-cloud.htm) * [Applications Services](https://docs.oracle.com/en-us/iaas/Content/applications-manager/applications-services-home.htm) * [Infrastructure Services](https://docs.oracle.com/en-us/iaas/Content/services.htm) * [Developer Resources](https://docs.oracle.com/en-us/iaas/Content/devtoolshome.htm) * [Security](https://docs.oracle.com/en-us/iaas/Content/Security/Concepts/security.htm) * [Marketplace](https://docs.oracle.com/en-us/iaas/Content/Marketplace/home.htm) * [Partner Offerings: Cloudflare at OCI](https://docs.oracle.com/en-us/iaas/Content/General/partner-offerings/cloudflare.htm) * [More Resources](https://docs.oracle.com/en-us/iaas/Content/General/Reference/more.htm) * [Glossary](https://docs.oracle.com/en-us/iaas/Content/libraries/glossary/glossary-intro.htm) * [Oracle Cloud Console](https://cloud.oracle.com/) [](https://docs.oracle.com/) Main Navigation * [Getting Started with Oracle Cloud Infrastructure](https://docs.oracle.com/iaas/Content/GSG/Concepts/baremetalintro.htm) * [Cloud Adoption Framework](https://docs.oracle.com/iaas/Content/cloud-adoption-framework/home.htm) * [Oracle Cloud's Free Tier](https://docs.oracle.com/iaas/Content/FreeTier/freetier.htm) * [Multicloud](https://docs.oracle.com/iaas/Content/multicloud/Oraclemulticloud.htm) * [Oracle Dedicated Cloud](https://docs.oracle.com/iaas/Content/dedicated/home.htm) * [Oracle Cloud Infrastructure Government Cloud](https://docs.oracle.com/iaas/Content/gov-cloud/govlanding.htm) * [Applications Services](https://docs.oracle.com/iaas/Content/applications-manager/applications-services-home.htm) * [Infrastructure Services](https://docs.oracle.com/iaas/Content/services.htm) * [Developer Resources](https://docs.oracle.com/iaas/Content/devtoolshome.htm) * [Security](https://docs.oracle.com/iaas/Content/Security/Concepts/security.htm) * [Marketplace](https://docs.oracle.com/iaas/Content/Marketplace/home.htm) * [More Resources](https://docs.oracle.com/iaas/Content/General/Reference/more.htm) * [Oracle Cloud Console](https://cloud.oracle.com/) * [Cloud Applications Documentation](https://docs.oracle.com/en/cloud/saas/index.html) * [Oracle.com Home](https://oracle.com/) * [Oracle Help Center Home](http://docs.oracle.com/) [Help Center](https://docs.oracle.com/) Search **Oracle Cloud Infrastructure** Search is scoped to: Oracle Cloud Infrastructure ### [Oracle Cloud Infrastructure Documentation](https://docs.oracle.com/iaas/Content/home.htm) * * * [Try Free Tier](https://www.oracle.com/cloud/free/?source=:ow:o:h:po:OHPPanel1nav0625&intcmp=:ow:o:h:po:OHPPanel1nav0625) All Pages * [Compatible MiniMax Models](https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm#top) * [MiniMax M3](https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm#minimax-m3) * [MiniMax M2](https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm#minimax-m2) Was this article helpful? Updated 2026-09-04 [Skip to main content](https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm#dcoc-content-body) Updated 2026-09-04 # Compatible MiniMax Models You can import large language models from Hugging Face and OCI Object Storage buckets into OCI Generative AI, create endpoints for those models, and use them in the Generative AI service. **Note** Before selecting a model, review the guidance following the model tables on this page for selecting a compatible hardware unit shape and understanding billing for dedicated AI clusters that host imported models. ## MiniMax M3🔗 The MiniMax-M3-MXFP8 model is the MXFP8 quantized variant of MiniMax M3, a native multimodal model with a one million token context. The model has about 428 billion total parameters with about 23 billion activated parameters and uses MiniMax Sparse Attention (MSA) for efficient long-context processing. This model is optimized for coding, long-horizon agentic workflows, and collaborative productivity tasks. Compatible MiniMax M3 Models| Hugging Face Model ID | Model Capability | Minimum Dedicated AI Cluster Unit Shape | | --- | --- | --- | | [MiniMaxAI/MiniMax-M3-MXFP8](https://huggingface.co/MiniMaxAI/MiniMax-M3-MXFP8) | TEXT_TO_TEXT | * H200_X8 * B200_X8 | ## MiniMax M2🔗 The MiniMax M2 text-to-text models are optimized for coding, complex reasoning, and agentic workflows such as tool use, search, and productivity tasks. MiniMax-M2 is a Mixture-of-Experts (MoE) model designed for efficient coding and agentic performance, and later MiniMax-M2 models extend this focus to more advanced software engineering and professional-work tasks. For more details, see MiniMax in the Hugging Face documentation. Compatible MiniMax M2 Models| Hugging Face Model ID | Model Capability | Minimum Dedicated AI Cluster Unit Shape | | --- | --- | --- | | [MiniMaxAI/MiniMax-M2.7](https://huggingface.co/MiniMaxAI/MiniMax-M2.7) | TEXT_TO_TEXT | * H100_X8 * H200_X8 | | [MiniMaxAI/MiniMax-M2.5](https://huggingface.co/MiniMaxAI/MiniMax-M2.5) | TEXT_TO_TEXT | * H100_X8 * H200_X8 | | [MiniMaxAI/MiniMax-M2](https://huggingface.co/MiniMaxAI/MiniMax-M2) | TEXT_TO_TEXT | * H100_X8 * H200_X8 | **Important** * For imported models, you can use the native context length specified by the model provider. However, the effective maximum context length is limited by the underlying hardware setup that you select for the hosting dedicated AI clusters in OCI Generative AI. To take full advantage of a model's native context length, you might need to provision more hardware resources. * Use the fine-tuned models only if they match the compatible base model's transformer version and have a parameter count within ±10% of the original. * To decide where you can host an imported model, find a compatible unit shape for the model in the table on this page, and then confirm that the hardware shape is available in the target region in [Hardware Unit Shapes by Region](https://docs.oracle.com/iaas/Content/generative-ai/hardware-unit-shapes-by-region.htm#imported-model-shapes). A model can be hosted in a region only when the region provides a hardware shape listed for that model. * For steps on how to deploy the imported models, see [Managing Imported Models](https://docs.oracle.com/iaas/Content/generative-ai/manage-imported-models.htm#imported-models "In addition to using the hosted pretrained models in OCI Generative AI, you can import validated open source and third-party models (for example, from Hugging Face) into OCI Generative AI, host them, create endpoints, and use them like any other model."). **Note** **Billing for Dedicated AI Clusters Hosting Imported Models** Dedicated AI clusters that host **imported models** don't require the 744-unit-hour minimum commitment that applies to dedicated AI clusters hosting pretrained models available in OCI Generative AI. Instead, an imported-model dedicated AI cluster has a minimum billable duration of one hour. For example, a cluster that uses one AI unit and runs for 30 minutes is billed for one AI unit-hour. After the first hour, charges are based on the cluster's actual running time and are prorated to the millisecond. For example, a cluster that runs for one hour and one minute is billed for that duration, rather than for two full hours. OCI sends usage records to Metering and Billing every five minutes while the cluster is running. If the cluster is deleted before the one-hour minimum is reached, OCI continues reporting billable usage until the minimum is met. If the cluster runs for more than one hour, OCI continues sending usage records every five minutes until the cluster is deleted. Was this article helpful? * [Oracle Cloud Infrastructure Cloud Adoption Framework](https://docs.oracle.com/en-us/iaas/Content/cloud-adoption-framework/home.htm) * [Getting Started](https://docs.oracle.com/en-us/iaas/Content/GSG/Concepts/baremetalintro.htm) * [Oracle Cloud's Free Tier](https://docs.oracle.com/en-us/iaas/Content/FreeTier/freetier.htm) * [Oracle Multicloud](https://docs.oracle.com/en-us/iaas/Content/multicloud/Oraclemulticloud.htm) * [Oracle Dedicated Cloud](https://docs.oracle.com/en-us/iaas/Content/dedicated/home.htm) * [Oracle Cloud Infrastructure Government Cloud](https://docs.oracle.com/en-us/iaas/Content/gov-cloud/govlanding.htm) * [Oracle UK Sovereign Cloud](https://docs.oracle.com/en-us/iaas/Content/uk-sovereign-cloud/home.htm) * [Oracle EU Sovereign Cloud](https://docs.oracle.com/en-us/iaas/Content/sovereign-cloud/eu-sovereign-cloud.htm) * [Applications Services](https://docs.oracle.com/en-us/iaas/Content/applications-manager/applications-services-home.htm) * [Infrastructure Services](https://docs.oracle.com/en-us/iaas/Content/services.htm) * [Developer Resources](https://docs.oracle.com/en-us/iaas/Content/devtoolshome.htm) * [Security](https://docs.oracle.com/en-us/iaas/Content/Security/Concepts/security.htm) * [Marketplace](https://docs.oracle.com/en-us/iaas/Content/Marketplace/home.htm) * [Partner Offerings: Cloudflare at OCI](https://docs.oracle.com/en-us/iaas/Content/General/partner-offerings/cloudflare.htm) * [More Resources](https://docs.oracle.com/en-us/iaas/Content/General/Reference/more.htm) * [Glossary](https://docs.oracle.com/en-us/iaas/Content/libraries/glossary/glossary-intro.htm) * [Oracle Cloud Console](https://cloud.oracle.com/) * Copyright © 2026, Oracle and/or its affiliates. * [About Oracle](https://www.oracle.com/corporate/index.html) * [Contact Us](https://www.oracle.com/corporate/contact/index.html) * [Legal Notices](https://docs.oracle.com/iaas/Content/legalnotices.htm) * [Terms of Use & Privacy](https://www.oracle.com/legal/privacy/) * [Document Conventions](https://docs.oracle.com/iaas/Content/General/Reference/docconventions.htm) * [Cookie Preferences](https://docs.oracle.com/iaas/Content/generative-ai/imported-minimax-models.htm)