# What is a chat model in n8n? > A chat model is the n8n sub-node that links an AI Agent to a language model. How to pick one, what drives the cost, and a receipt-reading example. Source: https://learn.ynteractive.com/content/glossary/chat-model · Updated 2026-10-01 · Free from Stepthrough (https://learn.ynteractive.com) [AI agents](https://learn.ynteractive.com/content/glossary#topic-ai-agents) · [AI glossary](https://learn.ynteractive.com/content/glossary) Updated October 1, 2026 · By [Robert Breen](https://learn.ynteractive.com/content/about) In n8n, a chat model is the sub-node that connects an AI Agent to a specific language model, such as one of OpenAI's, through your API credential. The agent decides what to do; the chat model is the part that actually reads and writes the text. ## Why it matters for a small business Picking the model is a business decision with three trade-offs: cost, speed and ability. A small, fast model is plenty for sorting form submissions or writing a two-line follow-up. A job that reads images or follows long, detailed rules may need a stronger one. Paying for the biggest model on every run adds up; a model that is too weak gives you sloppy rows and drafts you fix by hand. The chat model is also where the bill comes from. Every run sends your prompt, system message and recent memory to the provider, and the provider charges your API account for it. That account is separate from a ChatGPT subscription, and a key with no credit behind it simply returns errors. ## In a real lesson: AI Receipt Extractor: Receipts to Google Sheets with n8n The [AI Receipt Extractor lesson](https://learn.ynteractive.com/content/ai-receipt-extractor-n8n-google-sheets) shows why the choice matters. You click the **+** under **Chat Model** (the agent shows a red warning until you do), search **openai** and pick **OpenAI Chat Model**. In **Credential** you reuse **OpenAI account**, the key saved in an earlier module, so there is nothing new to paste. Then you open **Model** and change it to **gpt-4o**. The voice-over gives the reason: the default is fine for text, but this agent has to read a photo. In the test, it reads the Corner Mart receipt, picks Groceries on its own and writes the date, vendor and $31.57 total to the Invoices sheet. Compare the [AI Social Media Idea Agent](https://learn.ynteractive.com/content/n8n-ai-agent-tutorial-google-sheets), where you pick **gpt-5-mini** because generating text ideas is an easy job, and the email responder, which keeps the default model for short replies. [Try this lesson free](https://learn.ynteractive.com/modules/receipt-extractor) or [read the step-by-step guide](https://learn.ynteractive.com/content/ai-receipt-extractor-n8n-google-sheets). ## Common confusions ### Chat model vs ChatGPT ChatGPT is OpenAI's app with its own subscription. A chat model in n8n calls OpenAI's models through the [OpenAI API](https://learn.ynteractive.com/content/glossary/openai-api) with an API key, billed by usage on a separate platform account. Paying for one does not pay for the other. ### Chat model vs AI Agent The [AI Agent node](https://learn.ynteractive.com/content/glossary/ai-agent-node) holds the instructions and tools and decides the steps. The chat model only supplies the thinking. You can swap in a cheaper model or another provider without rebuilding the agent. ### "Chat" model vs a chat window The name refers to the message format the model works with, not to a chat screen. A chat model works the same behind a Gmail trigger or a webhook as behind the Chat Trigger. ## Tips - Start with a small, fast model and move up only when results are actually wrong. - Images need a model that accepts them. If an agent ignores an attachment, check the model before the prompt. - Add credit to the provider account before testing; an empty balance shows up as an error on the model sub-node. ## Related terms [AI model](https://learn.ynteractive.com/content/glossary/ai-model) · [AI Agent node](https://learn.ynteractive.com/content/glossary/ai-agent-node) · [OpenAI API](https://learn.ynteractive.com/content/glossary/openai-api) · [API usage billing](https://learn.ynteractive.com/content/glossary/api-usage-billing) · [Credential](https://learn.ynteractive.com/content/glossary/credential) · [Multimodal AI](https://learn.ynteractive.com/content/glossary/multimodal-ai) ## Where you use it: free lessons - [AI Receipt Extractor: Receipts to Google Sheets with n8n](https://learn.ynteractive.com/content/ai-receipt-extractor-n8n-google-sheets) (n8n, 12 min) - [n8n AI Agent Tutorial: Save Social Media Ideas to Google Sheets](https://learn.ynteractive.com/content/n8n-ai-agent-tutorial-google-sheets) (n8n, 12 min) - [AI Email Responder: Draft Gmail Replies Automatically with n8n](https://learn.ynteractive.com/content/ai-email-responder-gmail-drafts-n8n) (n8n, 12 min) ## Frequently asked questions **Which chat model should I pick in n8n?** The smallest one that does the job well. The lessons use gpt-5-mini for text ideas and form handling, and gpt-4o when the agent has to read a receipt photo. **Can I use a model other than OpenAI in n8n?** Yes. n8n has chat model sub-nodes for several providers, each with its own credential. The agent and its tools stay the same when you swap the model. **Why does my chat model fail with a brand-new API key?** A common cause is an API account with no credit. Add credit on the provider's billing page, then test again. [All AI glossary terms, A to Z](https://learn.ynteractive.com/content/glossary) · [Free prompt templates](https://learn.ynteractive.com/content/prompts)