GPT-6 Luna

Luna is the volume tier of GPT-6 — the model for jobs you run thousands of times a day. It keeps the flagship’s full 1.05M-token context and 128K-token answers, which is unusual at this price. At maximum effort OpenAI reports 66.6% on DeepSWE, which it puts level with Claude Opus 5 at medium effort.

Its reasoning can be switched off completely, so it also works as a fast plain model. Independent testing puts its intelligence roughly level with GPT-5.6 Luna: the gain is in price, not ability.

For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost $0.92 — 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.

GPT-6 Luna is billed in rupees, on the same monthly invoice as every other model your team uses — no foreign card, no separate OpenAI account, and one key for all of it.

Prompts sent to GPT-6 Luna 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.

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="gpt-6-luna",
    messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)