Google announced today that its Gemini 3.5 Live Translate model is now available to the public, marking a significant step toward seamless, real‑time multilingual conversation. The speech‑to‑speech system automatically detects spoken language, translates into more than 70 target languages, and reproduces the speaker’s intonation, pacing, and pitch. In practice, the model keeps pace with a normal dialogue, lagging only a few seconds behind the original utterance.
The rollout follows a series of staged demos that Google has showcased at its annual I/O event and other product launches. While earlier prototypes required specific hardware—such as Pixel phones or Pixel Buds—Google says the new offering works across a broader range of devices and integrates directly into the Translate app, Google Meet, and other services. The company frames the launch as the culmination of years of “pioneering machine‑learning experiments” aimed at breaking language barriers.
Gemini 3.5 Live Translate belongs to the 3.5 version family that debuted earlier this year. Until now, only the “Flash” variant of the model reached users, with Google hinting that a more powerful “Pro” version may arrive in the coming weeks. The current model processes speech continuously, handling multilingual inputs without requiring developers to configure language settings manually. It also includes built‑in noise suppression, allowing clear translation even in bustling environments.
Developers can begin building applications with the model through a public preview of the Gemini Live API and AI Studio. The API delivers a streamlined workflow: developers send a continuous audio stream, and the service returns translated speech in real time. Google emphasizes that the system’s low latency and natural‑sounding voice output reduce the robotic feel that has plagued earlier translation tools.
Industry observers note that the launch could accelerate adoption of real‑time translation in business, education, and travel. By embedding the technology across its ecosystem, Google positions itself to become the default platform for instant multilingual communication. The company has not disclosed pricing details for the upcoming Pro tier, but the preview suggests a free tier for developers experimenting with the technology.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.