Activefastmultimodal Proprietary

Gemini 3.1 Flash Live

by Google· Released May 2025

Gemini 3.1 Flash Live is a low-latency multimodal model from Google optimized for real-time interactions, including live audio and video streaming. It is part of the Gemini 3.1 Flash family, designed for fast, cost-effective performance with a 1M token context window.

Official Site API Docs

Input cost

$0.15 per 1M tokens

Output cost

$0.60 per 1M tokens

Context window

1M tokens

Max output

8192 tokens

Modalities

textimageaudiovideo

License

proprietary

Capabilities

Function CallingCode GenerationStreamingJSON ModeVisionAudio UnderstandingVideo UnderstandingLive Audio/Video

Best For

Real-time multimodal applications requiring low latency, such as live video analysis and interactive voice assistants.

Strengths

Ultra-low latency for live streaming
1M token context window
Multimodal (text, image, audio, video)
Cost-effective pricing

Limitations

Not suitable for complex reasoning tasks
May have lower accuracy on nuanced benchmarks compared to larger models
Limited to 8192 output tokens
Newer model with less community adoption

Use Cases

Real-time video captioning and analysis

Live customer support with audio/video

Interactive voice assistants

Live transcription and translation

Real-time content moderation

On-the-fly code assistance

Live educational tutoring

Improvements Over Previous Model

New model family; no direct predecessor in Flash Live variant
Optimized for live streaming with lower latency than Gemini 2.0 Flash
Supports live audio and video input natively
1M token context window (same as Gemini 2.0 Flash)
Pricing reduced compared to Gemini 2.0 Flash ($0.15 vs $0.35 per 1M input tokens)

Back to all models