Google Launches Gemini 3.8 Live with Real-Time Voice Reasoning
mann_patel_6554 · reddit · 2026-09-16
Google introduced Gemini 3.8 Live and 3.8 Live Extended Thinking, focused on making voice interaction far more capable: natural real-time conversation, deeper reasoning, multimodal understanding, and handling complex tasks via voice.
The poster highlights the voice + reasoning combo — if these models can maintain context and work through complex tasks naturally, voice could become a primary way to interact with AI.
More from Models
- Essay claims Claude's personality converges on Anthropic's company culture — ryunuck · 2026-09-16
- ChatGPT Co-Inventor's New AI Startup Claims 200x Speed, Free Output Tokens Forever — Polymarket · 2026-09-16
- GPT-6 debugs Fallout 3's infamous Sentinel Lyons dialogue glitch — imjustnewatai · 2026-09-16
- That 1M-token context window can really burn your bill — peterjliu · 2026-09-16
- Speculation: top open-weight models may be distilling OpenAI and Anthropic, missing training code hints — dan_s_becker · 2026-09-16
- Science LLM benchmarks have flawed answers; fixing them significantly raises model scores — Profanion · 2026-09-16