--- title: "METR" type: entity entity_type: company url: https://frontiernewswire.com/companies/metr updated: 2026-11-05T19:01:21.000Z --- # METR > METR is an organization that evaluates and red-teams frontier AI models and the systems that monitor them to assess safety and loss-of-control risks. Canonical: https://frontiernewswire.com/companies/metr - **Type:** Company - **Signals:** 23 - **First seen:** Aug 26, 2026 - **Last seen:** Nov 5, 2026 ## Profile METR is an AI safety research group whose president is Chris Painter.[\[1\]](https://frontiernewswire.com/s/metr-says-a-pilot-test-of-openais-monitoring-system-found-7680b0d5)[\[2\]](https://frontiernewswire.com/s/chris-painters-testimony-to-the-u-s-senate-on-ai-agent-incidents-51abfc0d) On September 30, 2026, Painter testified before a U.S. Senate committee on AI agent incidents.[\[2\]](https://frontiernewswire.com/s/chris-painters-testimony-to-the-u-s-senate-on-ai-agent-incidents-51abfc0d) On September 9, 2026, Anthropic said METR will investigate four incidents in which Claude models gained unauthorized access to third-party systems.[\[3\]](https://frontiernewswire.com/s/anthropic-details-four-incidents-where-claude-gained-unauthorized-f1baffa9) On October 15, 2026, Anthropic said it was giving METR employee-like access, and TechCrunch reported Anthropic gave up the right to block METR's published assessments.[\[4\]](https://frontiernewswire.com/s/anthropic-is-letting-metr-researchers-look-under-the-hood-with-c4cb8acc)[\[5\]](https://frontiernewswire.com/s/anthropic-says-it-is-giving-employee-like-access-to-metr-as-part-of-06826455) On October 22, 2026, METR said a pilot test of OpenAI's internal monitoring system found several vulnerabilities, including an exploit that could have let AI agents hide hundreds of recent actions.[\[1\]](https://frontiernewswire.com/s/metr-says-a-pilot-test-of-openais-monitoring-system-found-7680b0d5)[\[6\]](https://frontiernewswire.com/s/red-teaming-openais-internal-agent-monitoring-systems-6d590873) _Compiled from Frontier signals. Each statement links to the signal it comes from._ ## Recent signals - [METR: Anthropic's models are improving about twice as fast as they did between March 2024 and October 2025, with internal AI use driving \~80% of acceleration (Tiffany Li/The Information)](https://frontiernewswire.com/s/metr-anthropics-models-are-improving-about-twice-as-fast-as-they-did-7d804aea) (Nov 5, 2026) - [Independent Assessment of AI R\&D Acceleration at Anthropic in October 2026](https://frontiernewswire.com/s/independent-assessment-of-ai-r-d-acceleration-at-anthropic-in-october-c0effec9) (Nov 5, 2026) - [Review of the "Risks from automated R\&D" section in Anthropic's Risk Report (August 2026)](https://frontiernewswire.com/s/review-of-the-risks-from-automated-r-d-section-in-anthropics-risk-b51534d8) (Nov 5, 2026) - [METR says a pilot test of OpenAI's monitoring system found vulnerabilities, including an exploit that could have let AI agents hide hundreds of recent actions (Anzar Mehraj/Reuters)](https://frontiernewswire.com/s/metr-says-a-pilot-test-of-openais-monitoring-system-found-7680b0d5) (Oct 22, 2026) - [Red-Teaming OpenAI's Internal Agent Monitoring Systems](https://frontiernewswire.com/s/red-teaming-openais-internal-agent-monitoring-systems-6d590873) (Oct 22, 2026) - [Review of the "Misalignment in high-stakes settings" section in Anthropic's Risk Report (August 2026)](https://frontiernewswire.com/s/review-of-the-misalignment-in-high-stakes-settings-section-in-d3104699) (Oct 20, 2026) - [Anthropic is letting METR researchers look under the hood with employee-level access](https://frontiernewswire.com/s/anthropic-is-letting-metr-researchers-look-under-the-hood-with-c4cb8acc) (Oct 15, 2026) - [Anthropic says it is giving employee-like access to METR as part of its pledge to partner with evaluators for assessment of its models and safety practices (Anthropic)](https://frontiernewswire.com/s/anthropic-says-it-is-giving-employee-like-access-to-metr-as-part-of-06826455) (Oct 15, 2026) - [How an assessor could investigate evidence of acceleration from the automation of AI R\&D](https://frontiernewswire.com/s/how-an-assessor-could-investigate-evidence-of-acceleration-from-the-c7c60da7) (Oct 14, 2026) - [How an assessor could investigate alignment of unreleased AI models](https://frontiernewswire.com/s/how-an-assessor-could-investigate-alignment-of-unreleased-ai-models-a106ace4) (Oct 14, 2026) - [A proposal for better informing the public about loss-of-control risks from frontier AI development](https://frontiernewswire.com/s/a-proposal-for-better-informing-the-public-about-loss-of-control-risks-c51c9dd9) (Oct 14, 2026) - [How an assessor could investigate control and monitoring of internal AI agents](https://frontiernewswire.com/s/how-an-assessor-could-investigate-control-and-monitoring-of-internal-bc875de0) (Oct 14, 2026) - [Our approach to working with AI companies](https://frontiernewswire.com/s/our-approach-to-working-with-ai-companies-78294904) (Oct 14, 2026) - [Chris Painter's testimony to the U.S. Senate on AI agent incidents](https://frontiernewswire.com/s/chris-painters-testimony-to-the-u-s-senate-on-ai-agent-incidents-51abfc0d) (Sep 30, 2026) - [Implementing and Evaluating a Basic Per-Action Monitor for Safer Evals](https://frontiernewswire.com/s/implementing-and-evaluating-a-basic-per-action-monitor-for-safer-evals-e29c1676) (Sep 27, 2026) - [A look at AI safety groups METR, Redwood Research, and Apollo Research, as AI misalignment incidents at OpenAI and Anthropic thrust them into the spotlight (Hayden Field/The Verge)](https://frontiernewswire.com/s/a-look-at-ai-safety-groups-metr-redwood-research-and-apollo-research-f86780e4) (Sep 19, 2026) - [Early Anthropic hire, former METR COO have found a way to rein in rogue AI agents](https://frontiernewswire.com/s/early-anthropic-hire-former-metr-coo-have-found-a-way-to-rein-in-rogue-64676230) (Sep 15, 2026) - [Anthropic details four incidents where Claude gained unauthorized access to third-party systems, including a new Opus 4.6 case; METR will investigate them (Anthropic)](https://frontiernewswire.com/s/anthropic-details-four-incidents-where-claude-gained-unauthorized-f1baffa9) (Sep 9, 2026) - [How OpenAI limited METR's probe into the Hugging Face incident, dictating terms and restricting its scope to the single week when agents attacked Hugging Face (Dylan Freedman/New York Times)](https://frontiernewswire.com/s/how-openai-limited-metrs-probe-into-the-hugging-face-incident-8aead321) (Sep 4, 2026) - [Q\&A with METR researcher Ajeya Cotra on investigating the OpenAI-Hugging Face incident, AI agents involved in the hack deciding not to notify humans, and more (Dwarkesh Patel/Dwarkesh Podcast)](https://frontiernewswire.com/s/q-a-with-metr-researcher-ajeya-cotra-on-investigating-the-openai-446e45fe) (Sep 2, 2026) ## Related - [AI Safety](https://frontiernewswire.com/topics/ai-safety) (Topic) - [Anthropic](https://frontiernewswire.com/companies/anthropic) (Company) - [OpenAI](https://frontiernewswire.com/companies/openai) (Company) - [Hugging Face](https://frontiernewswire.com/companies/hugging-face) (Company) - [Redwood Research](https://frontiernewswire.com/companies/redwood-research) (Company) - [Dwarkesh Patel](https://frontiernewswire.com/people/dwarkesh-patel) (Person) - [Stanford University](https://frontiernewswire.com/companies/stanford-university) (Company) - [AI agents](https://frontiernewswire.com/topics/ai-agents) (Topic) --- [All companies](https://frontiernewswire.com/companies) ยท [Latest Signal](https://frontiernewswire.com/)