Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents
https://www.marktechpost.com/2026/09/15/google-releases-gemini-3-8-live-and-3-8-live-extended-thinking-for-production-grade-voice-agents/📌 Google發布Gemini 3.8 Live，語音Agent推理與說話同時進行
TL;DR：Google推出Gemini 3.8 Live與Extended Thinking，原生語音對語音模型讓voice agent邊思考邊回話，Speech to Speech Quality Index拿下全球第一。
語音agent長期以來卡在一個矛盾：要讓它一邊推理、一邊呼叫工具，又不能打斷對話的流暢度。Google這次發布的Gemini 3.8 Live系列，正是針對這道縫隙而來。
🤔 取代串接式語音管線的原生方案
Gemini 3.8 Live與Gemini 3.8 Live Extended Thinking是Google目前最先進的即時對話模型，屬於原生語音對語音（speech to speech）架構，是繼上月Gemini 3.5 Transcribe之後Gemini Audio家族的最新成員。Google將兩者定位為傳統ASR→LLM→TTS串接式語音管線的替代方案。兩款模型皆已在Gemini Live API與Google AI Studio上線，僅提供託管服務，沒有開放權重可自行部署。
🧩 兩款模型，各司其職
Gemini 3.8 Live鎖定規模化與成本效率，結合對話智能、流暢對答與視覺理解（visual grounding）。Gemini 3.8 Live Extended Thinking則針對高複雜度任務，在說話的同時加入多步驟推理能力。
Extended Thinking的核心設計是「邊想邊說」：模型會先用「讓我確認一下」之類的口頭提示回應使用者，再逐步敘述長任務執行的進度，背景則持續進行可配置的多步驟推理。Google展示的demo包括把手繪草圖搭配語音回饋轉換成可運作的React元件，以及協調多步驟的訂位／預約流程。Live API總共開放5項核心能力，Extended Thinking的可配置推理是其中之一。
📊 語音品質與agent任務雙雙拿下高分
| 基準 | 成績 |
|---|---|
| Speech to Speech Quality Index（Artificial Analysis） | 82.6，全球第一（Extended Thinking） |
| τ-Voice（agent任務完成率） | 68.6%（Extended Thinking） |
| τ-Voice-banking（Sierra） | 35.1%（Extended Thinking） |
| Big Bench Audio（音訊推理） | 97.7%（Extended Thinking） |
| Speech Agent Arena（人類偏好評測） | 第二名（Gemini 3.8 Live） |
在ServiceNow的EVA-Bench上，Google表示這兩款模型把複雜工作流的Pareto Frontier往前推進，在Gemini Enterprise Agent Platform的Live API上同時兼顧任務準確度與對話品質。
💡 定價與生態系整合
音訊輸入定價為每分鐘0.005美元，音訊輸出為每分鐘0.018美元，Google表示此估算基於每百萬輸入token 3美元、每百萬輸出token 12美元計算。開發者也可透過Agora、Fishjam、LangChain、LiveKit、Pipecat、Vercel、Vision Agents等Live API整合夥伴處理即時媒體串流基礎設施。Google並提到Salesforce、Genspark、Lumeris等合作方肯定了模型的低延遲、流暢度與工具呼叫能力，GitHub上也提供了範例應用程式。
🎯 實務啟示
對正在建置語音agent的工程團隊來說，Gemini 3.8 Live系列最大的意義在於把「推理」與「說話」整合進單一原生模型，省去串接ASR、LLM、TTS三個獨立元件的延遲與工程複雜度。若任務涉及多步驟工具呼叫或需要即時敘述進度（例如訂位、查詢、客服流程），Extended Thinking版本的τ-Voice與Big Bench Audio成績值得作為選型參考；但由於僅提供API託管服務、不開放權重，需要自行部署或資料不出境的場景仍需評估替代方案。
🔗 來源
- 標題：Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents
- 作者／機構：Asif Razzaq（MarkTechPost）
- 連結：https://www.marktechpost.com/2026/09/15/google-releases-gemini-3-8-live-and-3-8-live-extended-thinking-for-production-grade-voice-agents/
#GeminiLive #VoiceAI #SpeechToSpeech #ConversationalAI #GoogleAI #VoiceAgents #LLM #RealTimeAI #ExtendedThinking #AIProductRelease
原始資料 MarkTechPost · 收集於 2026-09-16
摘要原文
Google has introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, its most advanced live dialogue models to date. Both are native speech to speech models built for real time voice agents. They extend the Gemini Audio family that Google expanded last month with Gemini 3.5 Transcribe . The release targets a specific gap: voice agents that can reason and execute tools without breaking conversational flow. Is it deployable? Yes, for API based production use. Both models are live today in the Gemini Live API and Google AI Studio . They are hosted models, not open weights, so there is no self hosted option. The launch covers 2 models with distinct roles. Gemini 3.8 Live is built for scale and cost efficiency. It combines conversational intelligence with fluid dialogue and visual grounding. Gemini 3.8 Live Extended Thinking is built for high complexity tasks. It adds increased intelligence and multi step reasoning while it speaks. Google positions both as a streamlined alternative to cascaded speech pipelines that chain ASR, an LLM, and TTS. Gemini 3.8 Live Extended Thinking takes the #1 overall spot on Artificial Analysis’ Speech to Speech Quality Index with a score of 82.6. It leads agentic task completion with 68.6% on τ-Voice and 35.1% on Sierra’s τ-Voice-banking benchmark. It also scores 97.7% on Big Bench Audio, a reasoning benchmark for audio models. Gemini 3.8 Live secured second place in the Speech Agent Arena, a human preference evaluation. On ServiceNow’s EVA-Bench , Google reports that the models push the Pareto Frontier for complex workflows. They balance task accuracy with conversational quality, measured on the Live API on Gemini Enterprise Agent Platform. The Live API exposes 5 core capabilities in the new models: Extended Thinking adds configurable thinking for multi step reasoning in the background. It reasons and speaks simultaneously, using early verbal cues such as “Let me check that” to acknowledge prompts. It then narrates progress step by step while long running tasks execute. Google’s demos show the model converting sketches plus voice feedback into working React components and coordinating multi step bookings. Both models are priced at $0.005/min for audio input and $0.018/min for audio output. Google states this estimate is based on $3/1M input tokens and $12/1M output tokens. Developers can also build through Live API integration partners that handle real time media streaming infrastructure. These include Agora , Fishjam , LangChain , LiveKit , Pipecat , Vercel , and Vision Agents . Google is also partnering with Salesforce, Genspark, and Lumeris, which cite the models’ latency, fluidity, and tool calling. Example apps are available on GitHub . Check out the technical details and the developer post . All credit goes to the researcher of this project. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter . Wait! are you on telegram? now you can join us on telegram as well. Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us The post Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents appeared first on MarkTechPost .
由 claude-code/sonnet 自動生成