Glossary · AI
LLM
Also called: large language model, foundation model, GPT model
An LLM, or large language model, is a neural network trained on large amounts of text that generates text by repeatedly predicting the next token, which lets it summarize, classify, extract and draft.
For automation purposes, an LLM is a function: text (and sometimes images) in, text out. Everything else, from "AI summary" to "AI agent", is built on that.
What they're good at
- Reading unstructured input: emails, invoices, resumes, chat messages.
- Classifying and extracting: "is this a refund request?", "pull out the vendor and total".
- Drafting: replies, summaries, reports in a given tone.
- Following a format, especially when asked for structured output.
Where they fail
- Facts they weren't given. A model will produce a plausible answer rather than say "I don't know" unless told to. Ground it with the data it needs (RAG) and tell it to leave fields empty when unsure.
- Arithmetic and exact matching. Let ordinary steps add totals and compare IDs.
- Instructions hidden in input. Untrusted text can steer the model (prompt injection).
How they're billed
Providers charge per token, with input and output priced separately and bigger models costing more per token. The same task can differ in cost many times over between a small and a frontier model, which is why picking the model per step matters more than any other cost decision.
In Fluxo Flow
Fluxo Flow doesn't host or resell a model. You connect your own key from any of 14 providers (OpenAI, Anthropic, Google Gemini and more, including any OpenAI-compatible endpoint) and choose the model per AI step. Your provider bills you at its price with no markup, and the run trace shows tokens and estimated cost for each step.
Read next
Related terms
- TokenA token is the unit of text a language model reads and writes, usually a word, part of a word or a punctuation mark, and the unit AI providers use to price usage and set context limits.
- Context windowA context window is the maximum number of tokens a language model can take into account in a single request, counting the instructions, the input data and the reply it writes.
- BYOK (bring your own key)BYOK, or bring your own key, means a software product runs its AI features on an API key from your own AI provider account, so usage is billed to you directly by the provider instead of being resold through the product.
- Structured outputStructured output is a model response in a predefined, machine-readable format, typically a JSON object matching a schema, so the next steps of a workflow can use the fields directly instead of parsing prose.
- Low-codeLow-code describes development tools where most of an application or workflow is built visually, with small amounts of code (expressions, scripts or API calls) added where the visual tools run out.
- Lead routingLead routing is the process of automatically assigning each new lead to the right salesperson, team or nurture track based on rules such as territory, company size, product interest, score or round-robin rotation.