Panduan Penggunaan Model openai/gpt-oss-120b:free pada OpenRouter.ai via Python

# Buat virtual environment (opsional)
python -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate

# Instal library yang dibutuhkan
pip install --upgrade pip
pip install requests tqdm

# .env
OPENROUTER_API_KEY=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

pip install python-dotenv

import os
import json
import requests
from dotenv import load_dotenv

load_dotenv()  # Muat .env

API_URL = "https://openrouter.ai/api/v1/chat/completions"
API_KEY = os.getenv("OPENROUTER_API_KEY")

HEADERS = {
    "Authorization": f"Bearer {API_KEY}",
    "Content-Type": "application/json",
    "HTTP-Referer": "https://yourapp.example.com",   # Optional, untuk tracking
    "X-Title": "OpenRouter Python Demo"               # Optional, untuk monitoring
}

def chat_gpt_oss(prompt: str, temperature: float = 0.7, max_tokens: int = 512) -> str:
    """Kirim prompt ke model openai/gpt-oss-120b:free dan kembalikan teks hasil.
    
    Args:
        prompt (str): Teks yang ingin diproses.
        temperature (float): Nilai randomness, 0.0‑1.0.
        max_tokens (int): Batas token output.
    
    Returns:
        str: Jawaban model.
    """
    payload = {
        "model": "openai/gpt-oss-120b:free",
        "messages": [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": prompt}
        ],
        "temperature": temperature,
        "max_tokens": max_tokens,
        "stream": False
    }

    response = requests.post(API_URL, headers=HEADERS, json=payload, timeout=30)
    if response.status_code != 200:
        raise RuntimeError(f"API error {response.status_code}: {response.text}")
    data = response.json()
    # Struktur respons: {"choices": [{"message": {"content": "..."}}]}
    return data["choices"][0]["message"]["content"].strip()

# Contoh pemanggilan fungsi
if __name__ == "__main__":
    contoh_prompt = "Jelaskan perbedaan antara supervised learning dan unsupervised learning dalam 150 kata."
    jawaban = chat_gpt_oss(contoh_prompt)
    print(jawaban)

python gpt_oss_demo.py

import time

while True:
    user_input = input("Anda: ")
    if user_input.lower() in {"exit", "quit"}:
        print("Bot: Sampai jumpa!")
        break
    reply = chat_gpt_oss(user_input, temperature=0.6)
    print(f"Bot: {reply}\n")
    time.sleep(0.5)  # Hindari spam request

from tqdm import tqdm

articles = [
    "... teks artikel 1 ...",
    "... teks artikel 2 ...",
    # dst.
]

summaries = []
for text in tqdm(articles, desc="Summarizing"):
    prompt = f"Ringkas teks berikut dalam 100 kata: \n\n{text}"
    summary = chat_gpt_oss(prompt, temperature=0.3, max_tokens=150)
    summaries.append(summary)

# Simpan hasil ke file
with open("summaries.txt", "w", encoding="utf-8") as f:
    for i, s in enumerate(summaries, 1):
        f.write(f"=== Artikel {i} ===\n{s}\n\n")

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel

app = FastAPI()

class PromptRequest(BaseModel):
    prompt: str
    temperature: float = 0.7
    max_tokens: int = 512

@app.post("/generate")
async def generate(req: PromptRequest):
    try:
        result = chat_gpt_oss(req.prompt, req.temperature, req.max_tokens)
        return {"output": result}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

import time

def safe_chat(prompt, retries=3, backoff=2):
    for attempt in range(1, retries + 1):
        try:
            return chat_gpt_oss(prompt)
        except Exception as err:
            if attempt == retries:
                raise
            time.sleep(backoff ** attempt)