Gemini 3.8 Flash

Gemini 3.8 Flash is Google's September 2026 Flash release, the third in six weeks, and Google calls it its most intelligent workhorse model. It is built on Gemini 3.7 Flash rather than a new base, and the gains land where agents live: DeepSWE v1.1 rises from 65.3% to 73.7%, Terminal-bench 2.1 from 85.8% to 89.4%, the top score in Google's launch table, OSWorld-2.0 from 50.6% to 59.0%, and HLE-Verified reaches 54.9%. The design choice behind the numbers is that 3.8 Flash works harder: on complex tasks it takes extra reasoning steps and calls tools iteratively, so expect more tokens per task at higher effort levels. It keeps the 1M-token context, the 64K output ceiling, tunable effort levels, and text, image, audio and video input. Reach for it as the default for long-horizon coding agents, computer use and multi-step enterprise workflows where you want frontier-class results at Flash pricing.

Quick start

Use any OpenAI-compatible client. Just change base_url and your key.

from openai import OpenAI

client = OpenAI(
    api_key="tl-xxxxxxxxxxxxxxxxxxxxxxxx",
    base_url="https://api.thalam.ai/v1",
)

response = client.chat.completions.create(
    model="google/gemini-3.8-flash",
    messages=[
        {"role": "user", "content": "Hello from Thalam!"}
    ],
)

print(response.choices[0].message.content)