{
  "contentStorage": "postgres",
  "title": "Gemini 3.6 Flash · 关系",
  "description": "Google DeepMind 于 2026 年 7 月 21 日 发布的 Gemini 3.6 Flash。",
  "url": "/zh/works/gemini-3-6-flash/relation.html",
  "entity": {
    "id": "work_gemini_3_6_flash",
    "type": "work",
    "title": "Gemini 3.6 Flash",
    "slug": "gemini-3-6-flash",
    "summary": "Google DeepMind 于 2026 年 7 月 21 日 发布的 Gemini 3.6 Flash。",
    "status": "published",
    "categories": [
      "category-science-technology"
    ],
    "tags": [],
    "sources": [
      {
        "id": "s1",
        "title": "Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber",
        "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/"
      },
      {
        "id": "s2",
        "title": "Gemini 3.6 Flash | Gemini API",
        "url": "https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"
      },
      {
        "id": "s3",
        "title": "Gemini Developer API pricing",
        "url": "https://ai.google.dev/gemini-api/docs/pricing"
      },
      {
        "id": "s4",
        "title": "Gemini 3.6 Flash | Gemini Enterprise Agent Platform",
        "url": "https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/gemini/3-6-flash"
      }
    ],
    "createdAt": "2026-11-11T13:45:28+08:00",
    "updatedAt": "2026-11-11T13:45:28+08:00",
    "updatedBy": "user:admin",
    "revision": 1,
    "body": "## 发布背景\n\nGemini 3.6 Flash 是 [[organization_google_deepmind|Google DeepMind]] 于 2026 年 7 月 21 日 发布的 Gemini 模型[^s1]。与 [[work_gemini_3_5_flash_lite|Gemini 3.5 Flash-Lite]]、[[work_gemini_3_5_flash_cyber|Gemini 3.5 Flash Cyber]] 同日发布，定位为面向生产级智能体的编码与知识工作主力模型。[^s1]\n\n## 定价与获取\n\n2026年12月31日前标准定价为每百万 token 输入 0.75 美元、输出 3.75 美元，2027年1月1日起恢复为输入 1.50 美元、输出 7.50 美元[^s3]。API 标识为 gemini-3.6-flash[^s1]。\n\n## 性能表现\n\n以下均为 Google DeepMind 在发布公告与技术报告中公布的厂商自报成绩，尚未经独立复现：\n\n- DeepSWE：49%（厂商自报），高于 Gemini 3.5 Flash 的 37%（厂商自报）[^s1]。\n- MLE-Bench：63.9%（厂商自报），高于 Gemini 3.5 Flash 的 49.7%（厂商自报）[^s1]。\n- OSWorld-Verified：83.0%（厂商自报），高于 Gemini 3.5 Flash 的 78.4%（厂商自报）[^s1]。\n- GDPval-AA v2：1421 分（厂商自报），高于 Gemini 3.5 Flash（厂商自报）[^s1]。\n\n## 主要特点\n\n上下文窗口 1,048,576 token，最大输出 65,536 token[^s1]。\n\n官方称其输出 token 用量比 [[work_gemini_3_5_flash|Gemini 3.5 Flash]] 少 17%，在 DeepSWE 等个别基准上最多可少 65%（厂商自报）[^s1]。\n\nGoogle Cloud 企业平台标注其支持将于 2026 年 11 月 19 日 退役，但官方说明 Gemini API 在该日期后仍可继续使用[^s1]。\n\n## 定位\n\n它是 Gemini 3.5 Flash 的直接后继，作为 Flash 系列的工作主力模型，之后由 Gemini 3.7 Flash 接替。截至 2026 年 11 月，该模型仍在提供。",
    "url": "/works/gemini-3-6-flash.html",
    "lang": "zh",
    "sourceLang": "zh",
    "availableLangs": [
      "zh"
    ],
    "workKind": "ai_model"
  },
  "entities": null,
  "relations": [
    {
      "id": "relation_work_gemini_3_6_flash_published_by_organization_google_deepmind",
      "fromId": "work_gemini_3_6_flash",
      "relationType": "published_by",
      "toId": "organization_google_deepmind",
      "sourceRefs": [
        {
          "entityId": "work_gemini_3_6_flash",
          "sourceId": "s1"
        }
      ],
      "label": "出版方",
      "targetId": "organization_google_deepmind",
      "targetTitle": "Google DeepMind",
      "targetUrl": "/zh/organizations/google-deepmind.html",
      "incoming": false
    }
  ],
  "mentions": [
    {
      "fromId": "work_gemini_3_7_flash",
      "fromTitle": "Gemini 3.7 Flash",
      "fromUrl": "/zh/works/gemini-3-7-flash.html",
      "fromType": "work"
    }
  ],
  "relationGroups": [
    {
      "type": "published_by",
      "label": "出版方",
      "items": [
        {
          "id": "relation_work_gemini_3_6_flash_published_by_organization_google_deepmind",
          "fromId": "work_gemini_3_6_flash",
          "relationType": "published_by",
          "toId": "organization_google_deepmind",
          "sourceRefs": [
            {
              "entityId": "work_gemini_3_6_flash",
              "sourceId": "s1"
            }
          ],
          "label": "出版方",
          "targetId": "organization_google_deepmind",
          "targetTitle": "Google DeepMind",
          "targetUrl": "/zh/organizations/google-deepmind.html",
          "incoming": false
        }
      ]
    }
  ],
  "lang": "zh",
  "availableLangs": [
    "zh"
  ],
  "infoFields": [
    {
      "key": "work_kind",
      "label": "作品类型",
      "value": "人工智能模型"
    }
  ],
  "pagination": {
    "page": 1,
    "pageSize": 50,
    "total": 1,
    "pageCount": 1
  },
  "relationsTotal": 1,
  "mentionsTotal": 1
}