Gemini 3.8 Flash

Gemini 3.8 Flash is the third Flash release in six weeks and Google’s most capable so far. On the DeepSWE coding test it scored 73.7%, just under Claude Opus 5 and ahead of GPT-5.6 Sol, and Google reports wins over Opus 5 on finance and legal agent tests. It reads the same mix as Gemini Pro — text, images, audio, video and PDFs — in a million-token window.

Once it starts, it is fast: about 300 tokens a second in independent tests. Starting is the slow part — about 13 seconds to the first token, against a median of 3 — and it is wordy, so tasks can cost more than its rates suggest.

For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost $6.90 — a long prompt and a page of answer each. Every rate is published and stays the same from one request to the next. See every model.

Gemini 3.8 Flash is billed in rupees, on the same monthly invoice as every other model your team uses — no foreign card, no separate Google account, and one key for all of it.

Prompts sent to Gemini 3.8 Flash are processed outside Sri Lanka. If a workload has to stay in the country, Qwen3.8-27B (Colombo) runs on our own servers in Colombo.

Google’s model card records weaker safety behaviour outside English than the previous Flash, and Google publishes no Sinhala or Tamil results. If you serve Sinhala- or Tamil-speaking users, test with real messages and keep your own checks in place.

Use the OpenAI SDK you already have — point it at Roar AI and name the model. Read the docs.

from openai import OpenAI

client = OpenAI(base_url="https://api.roar-ai.com/v1", api_key="roar_live_…")

reply = client.chat.completions.create(
    model="gemini-3.8-flash",
    messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)