Gemini 3.8 Live

Gemini 3.8 Live is Google's latest entry in its real-time dialogue line, announced alongside a sibling variant called Gemini 3.8 Live Extended Thinking. According to third-party reporting of Google's announcement, the pair are positioned as the company's "most advanced live dialogue models yet," succeeding the Gemini 3.1 Flash Live release from March 2026. The new models are intended to make spoken interactions with AI feel more intuitive and intelligent, reflecting Google's ongoing push to refine conversational fluency and responsiveness in its Gemini family.

The Extended Thinking variant is tuned for higher-complexity tasks that require multi-step reasoning, and is being deployed behind several of Google's voice-driven experiences, including the Gemini Live assistant as well as the recently introduced Gmail Live conversational search and Keep Live note creation workflows. This division of labor lets the base Gemini 3.8 Live handle natural real-time dialogue while the Extended Thinking sibling tackles structured, reasoning-heavy interactions across Google's productivity surfaces, giving developers and end users a live model tier suited to both quick exchanges and more deliberate problem-solving.

Vercel AI Gatewaygoogle/gemini-3.8-livegemini

Open provider page Provider docs

Quick Info

Provider: Vercel AI Gateway
Model key: google/gemini-3.8-live
Release date: Sep 15, 2026
Last updated: Sep 15, 2026
Input modalities
Output modalities
Capabilities: Base catalog fields only

Cost

Input token cost: $0.75
Output token cost: $4.50

Limits

Output tokens: 0 tokens
Context window: 0 tokens

Latest news about Gemini 3.8 Live

Vercel AI Gateway

Coverage

Independent coverage of the September 15, 2026 launch of Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, noting that this follows the earlier 3.1 Flash Live release in March 2026. The article reports specific benchmarks cited from Google's announcement: the #1 overall spot on Artificial Analysis' Speech to Speec The Gemini 3.8 Live base model is described as offering "conversational intelligence with fluid dialogue and visual grounding," capable of processing visual inputs in near real time and executing tools and API calls in the background while conversation continues, including mid-conversation transitions between 97 suppor

Vercel AI Gateway

Official sourceOfficial

Google announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on September 15, 2026 (post updated September 17), described by the company as its "most advanced live dialogue models yet." The announcement was authored by principal engineer Tom Ouyang and member of technical staff Malini Jaganathan on behalf of Gemini 3.8 Live is positioned as the volume/default option for fluid conversational dialogue with visual grounding and language detection across 97 supported languages, while Gemini 3.8 Live Extended Thinking adds higher background reasoning and multi-step planning for complex tasks. The blog frames these as upgrades t

Vercel AI Gateway

Official sourceRelease Notes

The official Gemini API release notes for September 15, 2026 document the general availability of two new audio-to-audio models for the Live API: gemini-3.8-live (the default low-latency option for voice agents and real-time dialogue without reasoning delays) and gemini-3.8-live-extended-thinking (the high-reasoning va According to the changelog, gemini-3.8-live ships with interleaved reasoning, default asynchronous function calling, and full session client content updates, targeting most low-latency voice agent use cases. The Extended Thinking variant is explicitly recommended when higher background reasoning is required. The entry