Last refreshed 2026-09-17. Next refresh: weekly.
Why use Gemini 3.8 Live Extended Thinking on Google AI Studio?
Google AI Studio offers Gemini 3.8 Live Extended Thinking with pay-as-you-go pricing at $0.75/1M input tokens. Google AI Studio is a model prototyping environment and API access point for Gemini models, offering an inference playground for developers to test and build AI applications.
Setup recipe
Python + curlpip install google-genaiexport GOOGLE_API_KEY=...import os
from google import genai
client = genai.Client(api_key=os.environ["GOOGLE_API_KEY"])
response = client.models.generate_content(gemini-3.8-live-extended-thinkingRequest example
import os
from google import genai
client = genai.Client(api_key=os.environ["GOOGLE_API_KEY"])
response = client.models.generate_content(
model="gemini-3.8-live-extended-thinking",
contents="Hello"
)
print(response.text)Gotchas
- Use the model name directly, e.g. "gemini-2.0-flash", "gemini-1.5-pro", or "gemini-2.5-pro-preview-05-06".
- The examples expect GOOGLE_API_KEY; rename it only if your application config maps the new variable.
Pricing
| Type | Price (per 1M) |
|---|---|
| Input tokens | $0.75 |
| Output tokens | $4.50 |
| Audio input | $3.00 |
Capabilities
About Gemini 3.8 Live Extended Thinking
Gemini 3.8 Live Extended Thinking (API model code gemini-3.8-live-extended-thinking) is Google DeepMind's high-reasoning Live API audio-to-audio model for complex multi-step voice interactions, announced with Gemini 3.8 Live on September 15, 2026 (~17:00 UTC). First-party AI.dev: inputs text/images/audio/video; outputs text and audio; 131,072 input / 65,536 output tokens; Live API supported; Thinking supported; function calling async-only; search grounding supported; structured outputs / caching / code execution / batch not supported. Product blog positions it for deeper background reasoning while streaming continuous audio; ranks #1 on Artificial Analysis Speech-to-Speech leaderboard per Google blog claim.