Top news
Investing in the public benefit OfficialAnthropic NewsOct 18Popular Musk Pledges SpaceX Shares to ‘Every Kid in America’ WebBloomberg TechnologyOct 19Popular Websites are closing their doors to AI agents WebThe Verge AIOct 17 This is the new ChatGPT Voice, powered by GPT Live OfficialOpenAI on YouTubeOct 18 How AI giants will reshape global philanthropy WebAxios AIOct 19 Gemini 4 Argon Really Surprised Me WebTwo Minute PapersOct 19 Google Gemini 4 agents just beat Minecraft with no human intervention WebEngadget AIOct 19 Who could it be? 🧐 OfficialGoogle on YouTubeOct 18 Anthropic gives outside AI watchdog “employee-like” access WebArs Technica AIOct 18 What are self-replicating ‘AI worms’, and why is Anthropic worried about them? WebThe Guardian AIOct 18
GlossaryModelsIntermediate

Audio-to-audio

Also written: Native Audio

Audio-to-audio models take sound as input and produce sound as output, like a voice translator.

In plain terms

Audio-to-audio means a system understands spoken words and responds with spoken words. It takes sound in and creates sound out.

Think of a real-time voice translator. You speak in one language, and the system speaks back in another language instantly. This is different from systems that only understand speech and write it down (speech-to-text) or systems that only turn written text into speech (text-to-speech).

The expert version

Audio-to-audio models process audio input and generate audio output directly. This bypasses the need for intermediate text representations, potentially reducing latency and preserving nuances lost in text conversion. Applications include real-time translation, voice conversion, and expressive speech synthesis.

Why it matters now

Google Gemini's recent updates highlight audio-to-audio capabilities, particularly for real-time voice translation and conversational AI. These advancements focus on reducing latency and creating more natural voice interactions, suggesting a growing trend towards more integrated and immediate audio processing in AI systems.

Where it shows up 9 items, newest first
How each company uses it
Google Gemini
Google Gemini is actively developing and deploying audio-to-audio models for features like live translation and low-latency voice conversations.
Related terms

Speech recognitionSpeech synthesisVoice translationLow-latency AIReal-time AIMultimodal AI

Often confused with
Speech-to-text
Converts spoken audio into written text.
Text-to-speech
Converts written text into spoken audio.
Speech-to-speech translation
Translates spoken audio in one language to spoken
Origin

The concept of audio-to-audio processing in AI has evolved alongside advancements in speech recognition and synthesis. Direct audio-to-audio models represent a more integrated approach to spoken language tasks.

Try this about 5 minutes

Use a real-time voice translation app on your phone. Speak into the app in one language, and listen as it speaks your words in another language. This demonstrates how audio input is converted to audio output for immediate understanding.

Try it: Playbook entries that use thisPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

Questions people ask

What is the main advantage of audio-to-audio?
The main advantage is potentially lower latency and more natural-sounding output because the model processes audio directly without converting it to text first.
Can audio-to-audio models understand and speak any language?
Current audio-to-audio models are trained for specific language pairs or a set of languages. Their capabilities depend on the data they were trained on.

First seen on Subvolts Dec 12, 2025. Updated Oct 8, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.