What is a large language model?

A large language model is a neural network, almost always a transformer, trained on massive volumes of text with a single objective: given the tokens so far, predict the next one. At sufficient scale that one capability produces models that can summarise, translate, answer questions, classify text and write code.

What is the difference between an LLM and generative AI?

Generative AI is the broad category of models that produce new content, including images, audio and video. An LLM is the subset that works in text and code. Every LLM is generative AI, but generative AI also covers image and audio models that are not language models.

Do LLMs learn from the data I send them?

That depends entirely on your contract with the model provider, not on the model itself. Most enterprise API tiers exclude your inputs from training by default, and retention can usually be set to zero in your agreement with the provider. Under BYOM the contract is yours, so those terms are yours to set and to verify.

Which LLM should a Salesforce team use?

There is rarely one answer for a whole org. Long documents suit a model with a large context window, high-volume classification suits a small fast model, and data residency requirements may point to a self-hosted or regional deployment. Routing each workload separately usually beats standardising on a single model.

How does an LLM connect to Salesforce?

Through a Named Credential, which holds the provider endpoint and its authentication so credentials never appear in Apex or a Flow. GPTfy formats the prompt, masks sensitive fields, makes the callout under admin-controlled permissions, and writes the prompt and response to a Security Audit Record.

LLM (Large Language Model)

A neural network trained on massive text corpora to predict and generate text, the foundation behind ChatGPT, Claude, Gemini, and modern AI assistants.

Quick answer

What is LLM (Large Language Model)?

A large language model, or LLM, is a neural network trained on massive amounts of text to predict the next token, which lets it summarise, translate, answer questions and write code. In Salesforce, an LLM is the engine behind AI features, and under BYOM you choose which provider's model runs.

Last updated: August 2026

A large language model is a neural network, typically a transformer architecture, trained on hundreds of billions to trillions of tokens of internet text. The training objective is simple: given previous tokens, predict the next one. Scaled up sufficiently, this single capability produces models that can summarize, translate, answer questions, write code, and reason.

The frontier model families in production use come from OpenAI (GPT), Anthropic (Claude), Google (Gemini) and Meta (Llama), alongside a growing set of open-weight models that can be self-hosted. Version numbers move every few months, so enterprise teams generally choose on context window, latency, cost per token and measured performance on their own tasks rather than on brand or benchmark headlines.

For Salesforce: an LLM is the "AI engine" inside Agentforce, Einstein Copilot, and platforms like gptfy. The LLM doesn't store customer data. It processes prompts and returns responses. Architecture decisions about which LLM, how to send prompts, and how to handle responses define the entire AI experience.

Browse all terms

Frequently asked