Gemini 2.5 Flash Adds Native Real-Time Audio and Video Streaming for Developers

Google DeepMind released Gemini 2.5 Flash through the Gemini API and Google AI Studio, featuring built-in real-time audio and video streaming endpoints. The model targets sub-500ms first-token latency for voice applications and supports simultaneous multimodal inputs. Pricing is set at $0.075 per 1M input tokens, undercutting comparable multimodal API offerings.