●FLASH — Gemini 3.5 Flash is now generally available and powers gemini-flash-latest●AGENTS — Managed Agents in the Gemini API enter public preview, running autonomously in Google-hosted isolated Linux sandboxes●STUDIO — Google AI Studio adds project-level spend caps and native Android vibe coding●LIVE — Gemini 3.1 Flash Live Preview, an audio-to-audio model for real-time dialogue, is now available●IMAGE — Nano Banana 2 Lite arrives as the fastest, most cost-efficient image model, and Gemini Omni Flash is in API preview●LYRIA — Lyria 3 music generation models (clip-preview and pro-preview) are now available●FLASH — Gemini 3.5 Flash is now generally available and powers gemini-flash-latest●AGENTS — Managed Agents in the Gemini API enter public preview, running autonomously in Google-hosted isolated Linux sandboxes●STUDIO — Google AI Studio adds project-level spend caps and native Android vibe coding●LIVE — Gemini 3.1 Flash Live Preview, an audio-to-audio model for real-time dialogue, is now available●IMAGE — Nano Banana 2 Lite arrives as the fastest, most cost-efficient image model, and Gemini Omni Flash is in API preview●LYRIA — Lyria 3 music generation models (clip-preview and pro-preview) are now available