{"title":"Exercise: Making Your First LLM API Call","description":null,"content":"# Exercise 1: Send Your First Completion\n\nMake a single chat completion call through OpenRouter, print the model's answer, and print the token counts for the request.\n\nThis is the smallest complete unit of an AI application: send messages, get a response, account for the tokens you used. Everything else in the course builds on this call, so it is worth getting the shape of it into your fingers.\n\n### Requirements\n\n- Create a client pointed at OpenRouter.\n- Send a system message and a user message to `openai/gpt-5.4-mini`.\n- Print the model's reply.\n- Print the input, output, and total token counts from `response.usage`.\n\n\n**main.py**\n\n```python\nfrom openai import OpenAI\nimport os\nfrom dotenv import load_dotenv\n\nload_dotenv()\n\nclient = OpenAI(\n api_key=os.getenv(\"OPENROUTER_API_KEY\"),\n base_url=\"https://openrouter.ai/api/v1\",\n)\n\n# TODO: call client.chat.completions.create with a system and a user message\n\n# TODO: print the model's reply (response.choices[0].message.content)\n\n# TODO: print response.usage prompt_tokens, completion_tokens, and total_tokens\n```\n\n\n### Expected Output\n\nYou get a coherent answer followed by the token accounting. The wording of the answer varies between runs; the structure of the output does not.\n\n\n```plaintext\nIndexes speed up reads at the cost of slower writes and extra storage...\nTokens used - Input: 34, Output: 88, Total: 122\n```\n\n\n### Solution\n\n\n**main.py**\n\n```python\nfrom openai import OpenAI\nimport os\nfrom dotenv import load_dotenv\n\nload_dotenv()\n\nclient = OpenAI(\n api_key=os.getenv(\"OPENROUTER_API_KEY\"),\n base_url=\"https://openrouter.ai/api/v1\",\n)\n\nresponse = client.chat.completions.create(\n model=\"openai/gpt-5.4-mini\",\n messages=[\n {\"role\": \"system\", \"content\": \"You are a senior engineer. Answer concisely.\"},\n {\"role\": \"user\", \"content\": \"What does a database index do?\"},\n ],\n)\n\nprint(response.choices[0].message.content)\n\nprint(\n f\"Tokens used - Input: {response.usage.prompt_tokens}, \"\n f\"Output: {response.usage.completion_tokens}, \"\n f\"Total: {response.usage.total_tokens}\"\n)\n```\n\n\nThe only thing that makes this an OpenRouter call rather than a direct OpenAI call is the `base_url`. The `messages` list carries the system instruction and the user question, the reply lives at `response.choices[0].message.content`, and `response.usage` reports the exact token counts the provider billed, which is the number you log and budget against in a real application.\n\n---\n\n# Exercise 2: Hold a Two-Turn Conversation\n\nA single completion has no memory. To hold a conversation, you resend the whole message history each turn, including the model's own previous replies. Send a first message, append the reply, then ask a follow-up that can only be answered from the earlier turn.\n\n### Requirements\n\n- Send a first user message and print the reply.\n- Append the reply to `messages` as an `assistant` turn.\n- Append a follow-up user question that depends on the first message.\n- Send the updated history and print the second reply, which should recall the earlier detail.\n\n\n**main.py**\n\n```python\nfrom openai import OpenAI\nimport os\nfrom dotenv import load_dotenv\n\nload_dotenv()\n\nclient = OpenAI(\n api_key=os.getenv(\"OPENROUTER_API_KEY\"),\n base_url=\"https://openrouter.ai/api/v1\",\n)\n\nmessages = [{\"role\": \"user\", \"content\": \"My favorite language is Python.\"}]\n\n# TODO: send messages, print the reply, and append it as an assistant message\n\n# TODO: append a follow-up user message asking \"What did I say my favorite language was?\"\n\n# TODO: send again and print the reply, which should recall Python\n```\n\n\n### Expected Output\n\nThe second reply correctly recalls the detail from the first turn, proving the model saw the history.\n\n\n```plaintext\nTurn 1: Great choice, Python is versatile and beginner-friendly.\nTurn 2: You said your favorite language is Python.\n```\n\n\n### Solution\n\n\n**main.py**\n\n```python\nfrom openai import OpenAI\nimport os\nfrom dotenv import load_dotenv\n\nload_dotenv()\n\nclient = OpenAI(\n api_key=os.getenv(\"OPENROUTER_API_KEY\"),\n base_url=\"https://openrouter.ai/api/v1\",\n)\n\nmessages = [{\"role\": \"user\", \"content\": \"My favorite language is Python.\"}]\n\nfirst = client.chat.completions.create(model=\"openai/gpt-5.4-mini\", messages=messages)\nreply = first.choices[0].message.content\nprint(\"Turn 1:\", reply)\n\n# Carry the conversation forward by appending the assistant turn\nmessages.append({\"role\": \"assistant\", \"content\": reply})\nmessages.append({\"role\": \"user\", \"content\": \"What did I say my favorite language was?\"})\n\nsecond = client.chat.completions.create(model=\"openai/gpt-5.4-mini\", messages=messages)\nprint(\"Turn 2:\", second.choices[0].message.content)\n```\n\n\nThe model is stateless: it only knows what is in the `messages` you send. Appending the assistant's reply and then the follow-up rebuilds the full history, so the second call can answer a question about the first turn. Drop the assistant turn or the original message and the model loses the context, which is exactly why conversation management is about maintaining this list, not about the model remembering anything on its own.","pageType":"ai-engineering"}

Get Premium