Google rolls out Gemini 3.5 Pro and live‑voice translation for 70+ languages

Google has launched Gemini 3.5 Pro, a new AI model with a two‑million‑token context window—double that of its predecessor—and a Deep Think reasoning mode aimed at complex tasks. The Pro tier is priced at $20 per month, while access to Deep Think requires the $250 Ultra subscription. The release follows a rapid shift in the AI market, with ChatGPT’s share falling to roughly 53% and Gemini’s traffic rising to about 27%, supported by 900 million monthly Gemini app users.

In parallel, Google introduced Gemini 3.5 Live Translate, a real‑time speech translation system that preserves the original speaker’s intonation, tempo, and pitch across more than 70 languages. The feature is integrated into Google Translate, Google Meet, and an Android app that offers a “listening mode” for private audio playback. All AI‑generated audio carries a non‑audible SynthID watermark for detection. Both the Pro model and Live Translate are available to developers via the Gemini Live API and to Workspace customers in closed previews, signaling Google’s push to challenge OpenAI’s dominance in generative AI.