Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking

Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking at a glance

Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are both Google models in the Audio Generation category, available through the EmpirioLabs API.

Key differences

• Both models were released on Sep 15, 2026.
• Both models have a context window of 128K tokens.
• Capabilities listed only for Gemini 3.8 Live Extended Thinking: reasoning.
• Both models list the same rate for Input: audio (per 1M audio input tokens) and Input (per 1M prompt tokens).

Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking: specs

Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking: API pricing

Live rates from the EmpirioLabs catalog for both models. Each rate card always matches what the API charges.

Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking: API parameters

Request parameters each model lists on EmpirioLabs, with the allowed values or range where one applies. A dash marks a parameter the model does not list.

Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking: limits and notes

Gemini 3.8 Live

Connecting

Full-duplex voice over a WebSocket at wss://api.empiriolabs.ai/v1/realtime?model=gemini-3-8-live, authenticated with the ordinary Authorization Bearer header. Configure the session with session.update before you send the first audio: voice, instructions, turn detection and tools. These settings are fixed once the conversation starts.

Sending audio and video

• Append microphone audio as input_audio_buffer.append events: 16-bit PCM at 16 kHz, or at 24 kHz with input_audio_format set to pcm24.
• Append live video frames as input_image_buffer.append events, as base64 JPEG or PNG images.
• Voice activity detection is on by default, so keep streaming for about a second after the speech ends. Set turn_detection to null to end each turn yourself with input_audio_buffer.commit.

Voices and language

• Use one of the 30 voices in the voice list, for example Puck, Kore or Charon. Any other value is refused.
• The model replies in the language you speak.

Replies

• Speech streams as response.audio.delta events, 16-bit PCM at 24 kHz, with the words as response.audio_transcript.delta events. Your own speech is transcribed as well.
• Speaking while the model replies interrupts it.

Tools

• Function calling with JSON Schema parameters. Return each result as a function_call_output item with its call_id.

Billing

Audio and text tokens are priced separately in both directions, video frames bill as input tokens, and reasoning tokens bill as output tokens. Each completed turn is billed on its own from the usage the model reports, including a reply you interrupt.

Gemini 3.8 Live Extended Thinking

Connecting

Full-duplex voice over a WebSocket at wss://api.empiriolabs.ai/v1/realtime?model=gemini-3-8-live-extended-thinking, authenticated with the ordinary Authorization Bearer header. Configure the session with session.update before you send the first audio: voice, instructions, turn detection and tools. These settings are fixed once the conversation starts.

Sending audio and video

• Append microphone audio as input_audio_buffer.append events: 16-bit PCM at 16 kHz, or at 24 kHz with input_audio_format set to pcm24.
• Append live video frames as input_image_buffer.append events, as base64 JPEG or PNG images.
• Voice activity detection is on by default, so keep streaming for about a second after the speech ends. Set turn_detection to null to end each turn yourself with input_audio_buffer.commit.

Voices and language

• Use one of the 30 voices in the voice list, for example Puck, Kore or Charon. Any other value is refused.
• The model replies in the language you speak.

Replies

• Speech streams as response.audio.delta events, 16-bit PCM at 24 kHz, with the words as response.audio_transcript.delta events. Your own speech is transcribed as well.
• Speaking while the model replies interrupts it.

Reasoning

• reasoning_effort sets how much the model reasons in the background while it talks: low, medium (the default) or high. Set it before the conversation starts.
• The model often acknowledges a question first and answers in a second turn. Each turn is billed on its own.

Tools

• Function calling with JSON Schema parameters. Return each result as a function_call_output item with its call_id.

Billing

Audio and text tokens are priced separately in both directions, video frames bill as input tokens, and reasoning tokens bill as output tokens. Each completed turn is billed on its own from the usage the model reports, including a reply you interrupt.

How to call Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking

Gemini 3.8 Live holds a live conversation over a WebSocket at wss://api.empiriolabs.ai/v1/realtime?model=gemini-3-8-live, not over the HTTP endpoints. Connect with the model id gemini-3-8-live and send your EmpirioLabs API key as an Authorization: Bearer header on the handshake. Audio travels both ways as base64 16-bit PCM. A browser cannot set headers on a WebSocket, so open this connection from your server and relay audio to the browser over your own socket. Both models are called through the same endpoint. Set model to gemini-3-8-live or gemini-3-8-live-extended-thinking. Get an API key from the EmpirioLabs dashboard.

import asyncio, json, os, websockets

async def main():
    async with websockets.connect(
        "wss://api.empiriolabs.ai/v1/realtime?model=gemini-3-8-live",
        additional_headers=[
            ("Authorization", f"Bearer {os.environ['EMPIRIOLABS_API_KEY']}"),
        ],
        max_size=None,
    ) as ws:
        print(json.loads(await ws.recv())["type"])  # session.created

        await ws.send(json.dumps({
            "type": "conversation.item.create",
            "item": {
                "type": "message",
                "role": "user",
                "content": [{"type": "input_text", "text": "Say hello."}],
            },
        }))
        await ws.send(json.dumps({"type": "response.create"}))

        async for raw in ws:
            event = json.loads(raw)
            if event["type"] == "response.audio.delta":
                ...  # base64 audio chunk, append to your playback buffer
            elif event["type"] == "response.done":
                break

asyncio.run(main())

Gemini 3.8 Live vs Gemini 3.8 Live Extended Thinking: common questions

What is the difference between Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking?

Both models were released on Sep 15, 2026. Both models have a context window of 128K tokens. Capabilities listed only for Gemini 3.8 Live Extended Thinking: reasoning. The specs, rate cards, and parameters on this page come from the live EmpirioLabs catalog.

Which has the lower rate, Gemini 3.8 Live or Gemini 3.8 Live Extended Thinking?

Both models list the same rate for Input: audio (per 1M audio input tokens) and Input (per 1M prompt tokens). Both live rate cards are on this page and always match what the API charges.

Can I switch between Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking without rebuilding my integration?

Yes. Both models use WEBSOCKET /v1/realtime with the same EmpirioLabs API key, so most integrations switch by changing the model value. Check the parameters table on this page first if you rely on a model-specific parameter.

Can I try Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking in the browser before integrating?

Yes. Open each model in the EmpirioLabs playground and press Start session to try it in your browser.