Gemini 3.1 Pro

Gemini 3.1 Pro takes more kinds of input than any other model here: text, images, audio, video and PDFs, in a context of about a million tokens. Hand it an hour of meeting recording or a long scanned report and ask questions across all of it. It is a strong reasoner too — 77.1% on ARC-AGI-2 at launch, more than double Gemini 3 Pro — and scored 80.6% on SWE-Bench Verified.

Google still labels it a preview, eight months on, which means it can change under you. Its knowledge stops in January 2025, the oldest of the frontier models here, and the newer Gemini Flash releases now beat it on some agent tests.

For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost $20.70 — 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 3.1 Pro 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 3.1 Pro 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 publishes no Sinhala or Tamil test results for the Gemini API. If you plan to use it on recordings or documents in either language, try a sample before you commit.

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-3.1-pro",
    messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)