METR (Model Evaluation & Threat Research)
A nonprofit building the science of evaluating whether frontier AI systems could pose catastrophic risks.
To build the science of accurately assessing risks from frontier AI, so that humanity is informed before developing transformative AI systems. METR develops rigorous methods for measuring the autonomous and dangerous capabilities of frontier models and conducts independent, pre-deployment evaluations.
AI areas they serve
Areas of Focus
Frontier AI model evaluations
Autonomous and agentic capability assessment
Dangerous-capability and catastrophic-risk evaluation
Responsible Scaling Policies
AI R&D acceleration measurement
Pre-deployment evaluations for frontier labs
AI Impact Areas
Upcoming Goals
Advance the science of frontier AI evaluation and risk assessment
Conduct independent pre-deployment evaluations of frontier models
Prototype governance approaches tied to measured AI capabilities
Support standardized third-party AI evaluations
2025: Published influential research showing the length of tasks AI agents can complete has been doubling roughly every seven months
2023: Spun out from the Alignment Research Center as the independent nonprofit METR (renamed from ARC Evals)
2022: Founded by Beth Barnes as ARC Evals
All information on this page is from this Organization's website. All efforts are made to keep this information current.
If you represent this organization, please visit our Join Us page to provide any updates.