Gemini 2.5 Flash Live API native audio

Preview

This feature is subject to the "Pre-GA Offerings Terms" in the General Service Terms section of the Service Specific Terms. Pre-GA features are available "as is" and might have limited support. For more information, see the launch stage descriptions.

Gemini 2.5 Flash with Live API native audio features our cutting-edge native audio functionality for Live API. In addition to the standard Live API features, this preview model includes:

• Enhanced audio quality: Experience dramatically improved audio quality that feels like speaking with a person.
• Enhanced voice quality and adaptability: Live API native audio provides richer, more natural voice interactions with 30 HD voices in 24 languages.
• Introducing Proactive Audio: When Proactive Audio is enabled, the model only responds when it's relevant. The model generates text transcripts and audio responses proactively only for queries directed to the device, and does not respond to non-device directed queries.
• Introducing Affective Dialog: Models using Live API native audio can understand and respond appropriately to users' emotional expressions for more nuanced conversations.
• Improved barge-in: Interrupt Gemini more naturally and reliably, even in loud and noisy environments.
• Robust function calling: We've improved the triggering rate, allowing Gemini to successfully execute the functions you define to support your use cases.
• Accurate transcription: The accuracy of audio-to-text transcription has been significantly enhanced.
• Seamless multilingual support: Speak to Gemini in multiple languages, and it will effortlessly switch between them without any pre-configuration. Language is no longer a barrier.

For more information on Live API, see:

• Our standalone Live API documentation.
• Our Live API supported audio formats.
• Our Live API concurrent session limits.

Try in Vertex AI

Model ID	gemini-live-2.5-flash-preview-native-audio-09-2025
Supported inputs & outputs	Inputs: Text, Audio, Video Outputs: Text, Audio
Token limits	Maximum input tokens: 128K Maximum output tokens: 64K Context window: 32K (default), upgradable to 128K
Capabilities	Supported Grounding with Google Search System instructions Function calling Live API Preview feature Not supported Code execution Tuning Structured output Thinking Vertex AI RAG Engine Chat completions
Usage types	Supported Up to 1000 concurrent sessions Provisioned Throughput Not supported Dynamic shared quota Batch prediction
Technical specifications
	Video	Standard resolution: 768 x 768 Supported MIME types: video/x-flv, video/quicktime, video/mpeg, video/mpegs, video/mpg, video/mp4, video/webm, video/wmv, video/3gpp
	Audio	Maximum conversation length: Default 10 minutes that can be extended. Required audio input format: Raw 16-bit PCM audio at 16kHz, little-endian Required audio output format: Raw 16-bit PCM audio at 24kHz, little-endian Supported MIME types: audio/x-aac, audio/flac, audio/mp3, audio/m4a, audio/mpeg, audio/mpga, audio/mp4, audio/ogg, audio/pcm, audio/wav, audio/webm
	Parameter defaults	Start of speech sensitivity: Low End of speech sensitivity: High Prefix padding: 0 Max context size: 128K
Supported regions
	Model availability	United States us-central1
	See Data residency for more information.
Knowledge cutoff date	January 2025
Versions	gemini-live-2.5-flash-preview-native-audio-09-2025 Launch stage: Public preview Release date: September 18, 2025 gemini-live-2.5-flash-preview-native-audio Launch stage: Public preview Release date: June 17, 2025 Discontinuation date: October 18, 2025
Security controls
	See Security controls for more information.
Supported languages	See Supported languages.
Pricing	See Pricing.