- Founder and CEO, METR (Model Evaluation & Threat Research)12
Beth Barnes is the founder and CEO of METR (Model Evaluation and Threat Research), where Barnes oversees a technical team that designs and runs evaluations of generative AI models. Before METR, Barnes worked with DeepMind's Chief Scientist on scaling laws for forecasting deep learning progress, and at OpenAI on safety targets, scalable oversight techniques and pre-release evaluations of code models for misalignment.12
At a glance
- Current role
- Founder and CEO
- Company
- METR
- Sector
- Nonprofit
- Work listed
- 2
- Sources cited
- 5
- Profile reviewed
Career
Before founding METR, Barnes worked with DeepMind's Chief Scientist on scaling laws for forecasting progress in deep learning. At OpenAI, Barnes helped develop safety targets and evaluated scalable oversight techniques for alignment, as well as code models for misalignment before release.1
As founder and CEO of METR, Barnes leads an organization that evaluates frontier AI models to inform the public about their risks and capabilities, with a focus on whether AI systems can autonomously carry out substantial tasks.12
Research and evaluations
Barnes, listed on papers as Elizabeth Barnes, is a co-author of RE-Bench (2024), on evaluating language model agents' AI R&D capabilities against human experts, and a co-author of METR's 2025 study measuring AI ability to complete long software tasks. Barnes also co-authored METR's randomized controlled trial on how early-2025 AI tools affected experienced open-source developers' productivity.3451
Barnes's recent METR publications include external reviews of sections of Anthropic's risk reports in 2026 and work on a per-action monitor for safer evaluations.1
Work and contributions
RE-Bench Published research
Co-author of RE-Bench (2024), which compares language model agents against human experts on AI research and engineering tasks.3
Measuring AI Ability to Complete Long Software Tasks Published research
Co-author of METR's 2025 study measuring the length of software tasks that AI agents can complete.4
Timeline
- Co-authored RE-Bench.3
- Co-authored METR's study measuring AI ability to complete long software tasks.4
- Co-authored METR's randomized controlled trial on early-2025 AI and developer productivity.51
- Co-authored METR's write-up on implementing and evaluating a per-action monitor for safer evaluations.1
- Published METR's external review of a section of Anthropic's August 2026 Risk Report.1
Other documented roles
Roles the sources document that are not confirmed as current.
Where they work
- METR 5 people in the wiki
Related people
- Ajeya CotraAlso at METR
- Hjalmar WijkAlso at METR
- Nate RushAlso at METR
- Adam GleaveAlso at METR
- James GlasbrennerEvaluation and benchmarks · Safety and alignment
- Peter MattsonEvaluation and benchmarks · Safety and alignment
- Kush R. VarshneySafety and alignment · Evaluation and benchmarks
- Diogo AlmeidaEvaluation and benchmarks · Safety and alignment
- Jack ClarkSafety and alignment · Evaluation and benchmarks
- Jacob SteinhardtSafety and alignment · Evaluation and benchmarks
Suggest a correction or removal
Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.
References
5 sources · profile reviewed Oct 22, 2026 · data updated Oct 22, 2026
- Beth Barnes, Founder, CEO — METR team page
- About METR (leadership)
- RE-Bench: Evaluating frontier AI R&D capabilities of language model agents against human experts (arXiv 2411.15114)
- Measuring AI Ability to Complete Long Software Tasks (arXiv 2503.14499)
- Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity (arXiv 2507.09089)