Training
Open-weight Model
An AI model whose trained weights are publicly released, so anyone can download, run, or fine-tune it themselves.
01 ——
In plain English
An open-weight model is one where the trained model parameters (weights) are made publicly available — not just the code. You can download the weights, run the model on your own infrastructure, and fine-tune it for your needs without paying per-token API fees.
Examples:
- Meta's Llama family
- Mistral (Mistral 7B, Mixtral)
- Qwen (Alibaba)
- DeepSeek (R1, V3)
- Stable Diffusion (image)
Open-weight vs open-source: Strictly speaking, "open-source" requires the training code and data to also be public — most "open" AI models are open-weight but not fully open-source.
Why it matters:
- Privacy & control — keep data on your own servers
- Cost predictability — no per-token API charges
- Customisation — fine-tune freely
- Vendor independence — not locked to one provider
Open-weight models close the gap with frontier closed models every few months.
02 ——
Related terms
LLM
Large Language Model — the type of AI behind tools like ChatGPT and Claude, trained to understand and generate text.
Foundation Model
A large, general-purpose AI model trained on broad data that can be adapted (via prompting or fine-tuning) to many downstream tasks.
Fine-tuning
Further training a pre-trained AI model on your own data to specialise it for a specific task or style.
LoRA
Low-Rank Adaptation — a cheap way to fine-tune large AI models by training a small set of extra weights instead of the whole model.
Distillation
Training a smaller, cheaper AI model to mimic the outputs of a larger, more capable one — preserving most of the quality at a fraction of the cost.
Back to glossaryLast reviewed May 2026