{ "schemaVersion": "bitshovel-editorial-review-v1", "id": "reader-service-legacy-2-20260906", "signalId": "aihot-aihot-virxact-com-items-cmtacq8vz0aedroj24bcix9g", "originalUrl": "https://deepmind.google/blog/intelligent-transcription-with-gemini-3-5-transcribe", "kind": "evidence-review", "reviewedAt": "2026-09-06T04:01:40.570Z", "before": { "id": "aihot-aihot-virxact-com-items-cmtacq8vz0aedroj24bcix9g", "managedObjectId": "model-version:google-gemini-3-5-transcribe", "comparisonMember": "google-gemini", "state": "new", "lane": "productRadar", "publishedAt": "2026-08-26T17:01:00.000Z", "observedAt": "2026-08-27T00:42:11.669Z", "originalUrl": "https://deepmind.google/blog/intelligent-transcription-with-gemini-3-5-transcribe", "relatedSources": [ { "label": { "zh": "AIHOT 精选", "en": "AIHOT featured" }, "url": "https://aihot.virxact.com/items/cmtacq8vz0aedroj24bcix9go" }, { "label": { "zh": "The Verge 报道", "en": "The Verge coverage" }, "url": "https://www.theverge.com/tech/985186/google-gemini-3-5-transcribe-audio-ai" }, { "label": { "zh": "Ars Technica 报道", "en": "Ars Technica coverage" }, "url": "https://arstechnica.com/ai/2026/08/google-announces-gemini-3-5-transcribe-for-ai-powered-speech-to-text/" } ], "source": { "zh": "Google DeepMind 官方博客原文，经 AIHOT 精选进入视野", "en": "Google DeepMind blog via AIHOT featured" }, "title": { "zh": "Google 发布 Gemini 3.5 Transcribe：实时语音转写进入公开预览", "en": "Google ships Gemini 3.5 Transcribe into public preview" }, "summary": { "zh": "Google DeepMind 发布语音转文本模型 Gemini 3.5 Transcribe，官方称其为迄今最精确的转写模型：Artificial Analysis 测量的平均词错率为流式 4.0%、非流式 2.6%；自动检测并转写 85 种以上语言；对比上一代 Chirp 3，终稿转写用时缩短约 70%。模型以公开预览进入 Gemini API（AI Studio、Antigravity、Gemini Enterprise 可用），并已嵌入 Android Gboard 与 macOS 版 Gemini 应用。两条接口并行：Live API 流式版 gemini-3.5-transcribe-live 与处理预录音频的 gemini-3.5-transcribe（含说话人归属与词级时间戳）。", "en": "Google DeepMind launched Gemini 3.5 Transcribe, its self-described most precise speech-to-text model yet: Artificial Analysis measures average word error at 4.0% streaming and 2.6% non-streaming across 85+ auto-detected languages, and time to final transcription improves about 70% over Chirp 3. It ships as a public preview in the Gemini API and already powers Gboard on Android plus the macOS Gemini app. Two endpoints: streaming gemini-3.5-transcribe-live via the Live API, and gemini-3.5-transcribe for pre-recorded audio with speaker attribution and word-level timestamps." }, "why": { "zh": "前沿模型对比组同样缺 Google Gemini 成员的纵向锚点；这条还直接落在语音工作流上——85 语种覆盖加词级时间戳，对录音整理、客服质检和会议笔记类工具是一次明确的能力跳变。", "en": "The frontier-model lane also lacked a vertical anchor for Google Gemini; this lands squarely on voice workflows — 85-language coverage plus word-level timestamps is a clear capability jump for transcription, QA, and meeting-notes tooling." }, "caveat": { "zh": "全部指标出自 Google DeepMind 官方公告及其引用的 Artificial Analysis 测量；这不是独立评测，公开预览期的官方口径也不能证明生产环境下的稳定性或成本表现。", "en": "Every figure comes from Google DeepMind's own announcement and its cited Artificial Analysis measurements; this is not an independent evaluation, and official claims during a public preview are not proof of production stability or cost behavior." }, "timeline": [ { "date": { "zh": "8 月 27 日", "en": "Aug 27" }, "datetime": "2026-08-27T00:42:11.669Z", "summary": { "zh": "进入 BitShovel 雷达：AIHOT 精选 收录时可见 50 条。", "en": "Entered the BitShovel radar: 50 entries visible on AIHOT featured at collection." }, "source": { "zh": "AIHOT 精选 快照", "en": "AIHOT featured snapshot" } }, { "date": { "zh": "8 月 27 日", "en": "Aug 27" }, "datetime": "2026-08-27T00:42:11.669Z", "summary": { "zh": "The Verge 与 Ars Technica 同日跟进报道，交叉确认发布事实；两篇均强调去除语气词与自动排版等「智能转写」特性。", "en": "The Verge and Ars Technica ran same-day stories confirming the launch; both highlight the smart-transcription traits such as filler removal and automatic formatting." }, "source": { "zh": "二手媒体报道汇总", "en": "Secondary press roundup" } }, { "date": { "zh": "8 月 28 日", "en": "Aug 28" }, "datetime": "2026-08-27T22:06:27.440Z", "summary": { "zh": "市场横向（Google Gemini）：「Gemini Omni 1.1 Flash 发布,为开发者提供更强生成式视频控制」——AIHOT 精选 线索。同类的讨论与对比已入时间线，供纵向判断与淘汰决策参考。", "en": "Lateral (Google Gemini): \"Gemini Omni 1.1 Flash ships with stronger generative-video controls for developers\" (zh post, translated)" }, "source": { "zh": "市场横向 · 版本动态 · AIHOT 精选", "en": "Market lateral · Release watch · AIHOT featured" } }, { "date": { "zh": "8 月 28 日", "en": "Aug 28" }, "datetime": "2026-08-28T04:00:52.734Z", "summary": { "zh": "版本动态（Google Gemini）：「Gemini 3.5 Transcribe 发布:更精准的实时语音转写模型」——AIHOT 精选 线索。同类的版本动作，改变市场预期与替代选项。", "en": "Release watch (Google Gemini): \"Gemini 3.5 Transcribe arrives: a more precise real-time speech transcription model\" (zh post, translated)" }, "source": { "zh": "市场横向 · 版本动态 · AIHOT 精选", "en": "Market lateral · Release watch · AIHOT featured" } }, { "date": { "zh": "9 月 2 日", "en": "Sep 2" }, "datetime": "2026-09-01T22:00:14.205Z", "summary": { "zh": "版本动态（Google Gemini）：「Google DeepMind 为 Gemini 推出 agentic 视频理解功能」——AIHOT 精选 线索。同类的版本动作，改变市场预期与替代选项。", "en": "Release watch (Google Gemini): market chatter captured via AIHOT featured (headline in Chinese). Same-class release moves reshape expectations and alternatives." }, "source": { "zh": "市场横向 · 版本动态 · AIHOT 精选", "en": "Market lateral · Release watch · AIHOT featured" } }, { "date": { "zh": "9 月 3 日", "en": "Sep 3" }, "datetime": "2026-09-02T22:00:16.152Z", "summary": { "zh": "版本动态（Google Gemini）：「Google DeepMind 发布 Gemini 3.8 Flash 与 3.8 Flash 」——aihot-featured 线索。同类的版本动作，改变市场预期与替代选项。", "en": "Release watch (Google Gemini): market chatter captured via aihot-featured (headline in Chinese). Same-class release moves reshape expectations and alternatives." }, "source": { "zh": "市场横向 · 版本动态 · aihot-featured", "en": "Market lateral · Release watch · aihot-featured" } }, { "date": { "zh": "9 月 3 日", "en": "Sep 3" }, "datetime": "2026-09-02T22:00:16.152Z", "summary": { "zh": "对比评测（Google Gemini）：「Google DeepMind's Gemini 3.8 Flash Beats Claude 」——alphasignal-daily 线索。同类对比与口碑，是市场面相对位置的信号。", "en": "Comparison (Google Gemini): \"Google DeepMind's Gemini 3.8 Flash Beats Claude \" — lead via alphasignal-daily. Same-class comparisons and sentiment signal the market's relative standing." }, "source": { "zh": "市场横向 · 对比评测 · alphasignal-daily", "en": "Market lateral · Comparison · alphasignal-daily" } }, { "date": { "zh": "9 月 5 日", "en": "Sep 5" }, "datetime": "2026-09-05T01:00:21.319Z", "summary": { "zh": "版本动态（Google Gemini）：「谷歌 Gemini Spark 新增 Google Photos 照片库管理能力,用户可通过自然」——ithome 线索。同类的版本动作，改变市场预期与替代选项。", "en": "Release watch (Google Gemini): market chatter captured via ithome (headline in Chinese). Same-class release moves reshape expectations and alternatives." }, "source": { "zh": "市场横向 · 版本动态 · ithome", "en": "Market lateral · Release watch · ithome" } }, { "date": { "zh": "9 月 5 日", "en": "Sep 5" }, "datetime": "2026-09-05T10:00:24.514Z", "summary": { "zh": "版本动态（Google Gemini）：「Gemini 3.8 Flash is now available in GitHub Copi」——github-changelog 线索。同类的版本动作，改变市场预期与替代选项。", "en": "Release watch (Google Gemini): \"Gemini 3.8 Flash is now available in GitHub Copi\" — lead via github-changelog. Same-class release moves reshape expectations and alternatives." }, "source": { "zh": "市场横向 · 版本动态 · github-changelog", "en": "Market lateral · Release watch · github-changelog" } } ], "followTarget": { "id": "product:gemini-3-5-transcribe", "label": { "zh": "Gemini 3.5 Transcribe 产品线", "en": "Gemini 3.5 Transcribe product line" }, "scope": { "zh": "公测期到正式版的边界变化：GA 时间、定价、Chrome 听写上线与语种扩张", "en": "Preview-to-GA boundary shifts: general availability timing, pricing, Chrome dictation rollout, and language expansion" }, "url": "https://deepmind.google/blog/intelligent-transcription-with-gemini-3-5-transcribe", "updateMode": "verified-change" }, "ingestion": { "mode": "editorial", "adapter": "syndication-feed", "evidenceClass": "official-statement", "observationRef": "source-observations/history/syndication-feeds.2026-08-27T00-42-11Z.snapshot.json#feed:aihot-featured", "changeRef": "source-observations/history/syndication-feed-changes.2026-08-27T00-42-11Z.snapshot.json#feed:aihot-featured" }, "verifiedAt": "2026-09-05T16:05:43.235Z", "banner": { "url": "/card-art/gemini-transcribe-official.jpg", "credit": "Google DeepMind", "license": "© Google DeepMind - editorial use with attribution" } }, "after": { "title": { "zh": "Gemini 3.5 Transcribe：语音转写进入公开预览", "en": "Gemini 3.5 Transcribe enters public preview" }, "summary": { "zh": "Google 提供支持 85 种以上语言的流式与预录音频转写接口；预录接口可返回词级时间戳与说话人标签。发布公告列出的 macOS 应用当时仅支持英语，应用端可用范围需要与 API 能力分别查看。", "en": "Google offers streaming and prerecorded transcription across 85+ languages, with word timestamps and speaker labels for prerecorded audio. The launch announcement listed the macOS app as English-only; app availability should be checked separately from API capabilities." }, "why": { "zh": "开发访谈整理、字幕或会议工具时，时间戳有助于回到对应录音，核对说话人和原话。能否减少中文录音的整理时间，需要用自己的录音样本与现有工具比较。", "en": "For interview, subtitle, and meeting-tool developers, timestamps help readers return to the recording and check who said what. Test representative recordings against existing tools to assess the benefit for Chinese transcription." }, "caveat": { "zh": "使用时间戳或说话人区分时，预录音频上限为 30 分钟，且不能并用智能整理模式或自定义词表；3 人及以上的说话人归因仍属实验支持。本站尚未测试中文嘈杂录音的准确率。", "en": "With timestamps or speaker labels, prerecorded audio is limited to 30 minutes and cannot use smart mode or custom vocabulary. Attribution for three or more speakers remains experimental. We have not tested accuracy on noisy Chinese recordings." } }, "editorialReadiness": "decision-brief", "withdrawnSources": [], "withdrawnTimeline": [], "evidence": [ { "type": "retained-official-page", "label": { "zh": "blog.google 原文", "en": "Source page: blog.google" }, "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5-transcribe/", "path": "app/_content/editorial-reviews/evidence/reader-service-gemini-announcement-20260906.json", "bytes": 641, "sha256": "3bf3c584d9d8bd30a72a41292790638b1b1edb657974cb87d1cac512835352c3", "fetchedAt": "2026-09-06T03:49:20.293077+00:00" }, { "type": "retained-official-page", "label": { "zh": "ai.google.dev 原文", "en": "Source page: ai.google.dev" }, "url": "https://ai.google.dev/gemini-api/docs/transcribe", "path": "app/_content/editorial-reviews/evidence/reader-service-gemini-api-transcribe-20260906.json", "bytes": 551, "sha256": "2d35969cd1b1eb42fa0f6d0d307fd5d9b2f763f29a265b4a49c431b6581a7ff6", "fetchedAt": "2026-09-06T03:50:44.534993+00:00" } ], "sourceAssessments": [ { "url": "https://deepmind.google/blog/intelligent-transcription-with-gemini-3-5-transcribe", "role": "direct-support", "scope": { "zh": "原始 DeepMind 地址转到 Google 官方公告。当前 API 文档列出最多 8 位说话人，公告仍写最多 3 位；上文限制采用当前 API 文档，不据此推断变更发生日期。保留原条目 publishedAt=2026-08-26T17:01:00.000Z；9 月 6 日为补核验日期。本站未复测，不等于否认公告引用的 Artificial Analysis 测量的独立性。抓取元数据见 evidence-manifest.json。", "en": "The original DeepMind URL redirects to Google’s announcement. Current API documentation lists up to eight speakers; the announcement still lists three. Use current API limits without inferring when they changed. Retain publishedAt=2026-08-26T17:01:00.000Z; September 6 is verification only. Lack of our own testing does not negate the independence of cited Artificial Analysis measurements. See evidence-manifest.json." }, "limitations": { "zh": "使用时间戳或说话人区分时，预录音频上限为 30 分钟，且不能并用智能整理模式或自定义词表；3 人及以上的说话人归因仍属实验支持。本站尚未测试中文嘈杂录音的准确率。", "en": "With timestamps or speaker labels, prerecorded audio is limited to 30 minutes and cannot use smart mode or custom vocabulary. Attribution for three or more speakers remains experimental. We have not tested accuracy on noisy Chinese recordings." }, "evidenceUrls": [ "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5-transcribe/", "https://ai.google.dev/gemini-api/docs/transcribe" ] } ], "note": { "zh": "补充核对来源正文，更新用途、规格和使用条件。原发布日期与收录时间保留；本次编辑核验日期不代表产品刚发生变化。", "en": "Checked source materials to update purpose, specifications and usage conditions. Original publication and collection dates are retained; this review does not establish a new product event." }, "integrity": "66fc8fb7c9d8b3507aa92087d371047fbca919d208ac62316570f4dcfe9bf7c8" }