What is OpenAI?
After the Generative AI intro, we focus on one provider: OpenAI. You have probably used ChatGPT already. This page explains how that differs from the developer platform — no code in this lesson.
The company
OpenAI is a US research company, founded in 2015, that builds language and image models. Their chat product is ChatGPT. Developers who want models inside their own apps use the OpenAI Platform and pay per usage.
Later in this course, LangChain will sit in front of OpenAI (and other providers) so you can switch models without rebuilding your whole setup.
ChatGPT vs the API
Two entry points, same family of models underneath:
| ChatGPT | API | |
|---|---|---|
| Where | chat.openai.com | platform.openai.com |
| Typical use | Typing questions in a browser | Apps and services built by developers |
| Cost | Free tier + paid plans | Pay per token (pricing page) |
Common examples:
- Ganga (Hooghly in the lower reach)
- Damodar
- Teesta
Model names you will see
OpenAI renames models often. In ChatGPT you pick a model from a dropdown; on the API you pass the model name with each request. Two names worth knowing:
- GPT-4o — stronger, slower, costs more per request
- GPT-4o mini — cheaper, fine for everyday questions and drafts
For browsing and study help, the free ChatGPT tier is enough. The full model list is on the platform docs.
API keys (for later)
ChatGPT in the browser does not need an API key. The developer API does — you create one at platform.openai.com/api-keys when you are ready to build an app. We will use keys in a later lesson; for now, just know where they live.
API keys
| Name | Status | Tracking ID | Secret Key | Created | Last used ⓘ | Created by | Permissions | |
|---|---|---|---|---|---|---|---|---|
| langchain-course-dev | ✓ Active | key_0o56Q93L6xg78QCr… | sk-…Na8A | 7 Jun 2026 | Never | Qlist | All | ✎🗑 |
Treat a key like a password. If it leaks, revoke it from this page and create a new one.
What's Next
Claude, Gemini, Mistral, Llama, and when to use a cloud API vs a local model.