Models & architecture
Open-weights model
Definition
An open-weights model has publicly downloadable parameters, letting anyone run it on their own hardware. Llama, Mistral, Qwen and DeepSeek are prominent examples.
Open weights change the deployment calculus. You can run inference locally or in your own cloud, which means no per-token cost, no data leaving your infrastructure, no rate limits, and no vendor deprecating the model you built on.
The distinction from "open source" matters. Most open-weights releases publish parameters but not training data or training code, and many carry licences with commercial restrictions. Read the licence before building a business on one.
The capability gap against frontier closed models has narrowed considerably but not closed, particularly on complex reasoning. For many production tasks — classification, extraction, summarisation, routine generation — open weights are now entirely sufficient.
Related terms
Parameters
Parameters are the learned numerical weights inside a model, adjusted during training. Parameter count — 7B, 70B, 405B — is a rough proxy for capacity, but not a reliable predictor of quality.
Quantization
Quantization reduces the numerical precision of model weights — from 16-bit to 8-bit or 4-bit — shrinking memory use and speeding inference at a modest cost in quality.
Fine-tuning
Fine-tuning continues training a pre-trained model on your own examples to specialise its behaviour. It is the right tool for teaching consistent style, format or classification behaviour — not for adding knowledge.
Inference cost
Inference cost is what you pay to run a model on a request, almost always billed per million input and output tokens. Output tokens typically cost three to five times more than input tokens.
Large language model (LLM)
A large language model is a neural network trained on vast amounts of text to predict the next token in a sequence. That single objective, at sufficient scale, produces the ability to write, summarise, translate, reason and code.
Transformer
The transformer is the neural network architecture behind virtually all modern language models. Its key innovation is self-attention, which lets every token directly consider every other token in the input.
Put this into practice
Understanding the term is step one. Our free courses and tools let you actually use it.