Parameters: Parameters are the adjustable numbers inside an AI model, mainly its weights, that are learned during training. The parameter count, from a few billion to over a trillion, is the standard shorthand for a model's size and rough capability.
When you read that a model has 8 billion or 70 billion parameters, that is a count of the learned values that shape its behavior. More parameters give a model more capacity to store patterns and handle nuance, which is why the largest models tend to be the most capable at open-ended tasks.
Size comes with costs. A bigger model needs more memory and more computation for every response, so it is slower and more expensive to run. Small models with a few billion parameters can run on a laptop or phone and are often good enough for narrow tasks like classification, extraction, or drafting short replies. Vendors now offer families of models at several sizes for exactly this reason.
Parameter count is not the whole story. Training data quality, training method, and post-training matter enormously, and a well-trained smaller model can beat a larger, sloppier one on many tasks. Treat the number as one signal, then judge a model by how it performs on your actual work.
Example at work
An IT manager choosing a model for an internal document-tagging tool tests a small open-weights model and a large proprietary one on 200 real documents. The small model matches the large one on this narrow task at a fraction of the cost per run, so the team deploys it and reserves the large model for complex analysis.
Why it matters
Parameters are the most quoted number in AI marketing, so it helps to know what it does and does not tell you. The right question is rarely 'which model is biggest?' but 'which model is good enough, fast enough, and cheap enough for this job?'
Related terms
- Model weightsModel weights are the numbers inside a trained neural network that determine how it responds to input. They are what training produces and what a company ships or protects; releasing them publicly is what makes a model 'open weights.'
- Large language model (LLM)A large language model (LLM) is a neural network trained on enormous amounts of text to predict the next word, then tuned to follow instructions and hold a conversation. LLMs power ChatGPT, Claude, Gemini, and Microsoft Copilot.
- Neural networkA neural network is a computing system made of layers of simple connected units that pass numbers to each other, with connection strengths (weights) adjusted during training until the network produces useful output. It is the building block of modern AI.
- InferenceInference is the act of running a trained AI model to produce an output, such as answering a prompt or classifying an image. Training happens once; inference happens every time anyone uses the model, and it is what you pay for as a user.
- BenchmarkA benchmark is a standardized test used to measure and compare AI models on a specific skill, such as math, coding, knowledge, or reasoning, by scoring their answers against a fixed set of questions.