open-weight models

Open-weight models are artificial intelligence models whose weights, or parameters, are publicly available, allowing for broader access and modification compared to closed-source models.

Why it’s in the news: They are in the news due to their increasing adoption by major corporations and governments, sparking debates over their safety and regulation.

Latest on open-weight models

• Hong Kong firm bets on Chinese open-weight models to rival CoreWeave
• Why Capital One built its multi-agent AI platform around open-weight models
• Trump advisers tell AI firms they will not safety-test open-weight models
• Trump advisers tell AI firms open-weight models will not be put through safety tests
• US AI leaders turn to Chinese open-weight models, challenging closed-source safety claims
• Anthropic’s Dario Amodei responds: doesn’t oppose open-weight models, but fears Chinese AI
• open-weight models are the new new red scare
• Safety testing was supposed to prevent AI breaches
• Trump advisers told AI firms that open-weight models will not undergo safety testing
• Open-weight models distribute control across thousands of developers and organizations, each with different security postures and incentives. That fragmentation is the actual risk surface.
• Open-weight models democratize access, which sounds like victory until you realize it also democratizes deployment without governance
• When commodity models become good enough, the marginal value of training your own collapses overnight

Connections

11 entities linked to open-weight models across the news graph.