---
格式版本: 2
标题: "Compatible DeepSeek Models"
原文链接: "https://docs.oracle.com/en-us/iaas/Content/generative-ai/imported-deepseek-models.htm"
发布日期: "2026-08-20"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:scrape:strict_html_body"
发布时间证据: "cpp-last-modified-time: 2026-08-20 21:14:54Z"
发布时间校准原因: "规则确认唯一严格发布时间，来源 scrape:strict_html_body"
发布时间校准置信度: "high"
发布时间候选数量: 6
发布时间严格候选数量: 2
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: "2026-08-21T09:58:10+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: "2026-08-21T09:56:49+08:00"
入库时间: "2026-08-21T01:58:10.488Z"
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"
相关厂家:
  - "Oracle"
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 24
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "内容为OCI导入DeepSeek模型的部署文档，仅提及Dedicated AI Cluster Unit形状，无超节点/AI Rack架构、供电、散热或互连等技术细节，与项目关注硬件基础设施无关。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-21T09:58:22+08:00"
AI主题相关性: 1
AI来源权威性: 15
AI新颖性: 5
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 3
AI摘要: "Oracle OCI Generative AI 支持从 Hugging Face 导入多款 DeepSeek 模型，包括 DeepSeek-V4-Pro、V4-Flash 和 R1-Distill-Qwen-32B，创建端点后即可使用。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T02:13:19.161Z"
采集批次: "2026年8月21日9点50分40秒"
采集批次ID: "20260821-095040-035"
去重键: "https://docs.oracle.com/en-us/iaas/Content/generative-ai/imported-deepseek-models.htm"
---

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.

## DeepSeek

| Hugging Face Model ID | Model Capability | Minimum Dedicated AI Cluster Unit Shape |
| --- | --- | --- |
| [deepseek-ai/DeepSeek-V4-Pro](https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro) | TEXT\_TO\_TEXT |  |
| [deepseek-ai/DeepSeek-V4-Flash](https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash) | TEXT\_TO\_TEXT |  |
| [deepseek-ai/DeepSeek-V4-Flash-0731](https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731) | TEXT\_TO\_TEXT | - H100\_X4 - H200\_X4 - B200\_X2 - B300\_X2 |
| [deepseek-ai/DeepSeek-R1-Distill-Qwen-32B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B) | TEXT\_TO\_TEXT | A100\_80G\_X2 |

**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.
- For available hardware and 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.").