AIRiskAware
AI Governance Glossary
Governance Practice

What Is Open-Weight Model?

Open-Weight Model is an AI model whose trained parameters (weights) are publicly released, allowing anyone to run, fine-tune, or build on it.

Definition

Open-Weight Modelan AI model whose trained parameters (weights) are publicly released, allowing anyone to run, fine-tune, or build on it.

Open-weight models are sometimes loosely called "open source", but the terms differ: releasing weights is not the same as releasing the training data and code needed to reproduce the model. Open weights bring governance trade-offs — they enable scrutiny, customisation, and reduced vendor lock-in, but also make it harder to enforce safety guardrails or restrict misuse once the weights are distributed. This tension is central to current policy debates about how to regulate the most capable models.

Source: NIST AI 100-1; OECD AI Principles

Plain-language explanation

Open-weight models are sometimes loosely called "open source", but the terms differ: releasing weights is not the same as releasing the training data and code needed to reproduce the model. Open weights bring governance trade-offs — they enable scrutiny, customisation, and reduced vendor lock-in, but also make it harder to enforce safety guardrails or restrict misuse once the weights are distributed. This tension is central to current policy debates about how to regulate the most capable models.

Primary source: NIST AI 100-1; OECD AI Principles

See where you stand on AI governance

Take the free 7-question maturity assessment and get a personalised action plan.

Free assessment — 3 minutes →