Is this your company? Get featured or correct your listing →
Company Snapshot
METR is a AI infra company, based in Berkeley, California, USA.
Latest Pulse
- Oct 15, 2026Anthropic Gives AI Safety Group Employee-Level AccessAnthropic is giving six researchers from AI-safety nonprofit METR company laptops, office access and internal tools on par with its own risk teams, and has given up the right to block their findings from publication.→
- Sep 1, 2026Someone Stole an AI Safety Lab's API Key. Nobody Noticed.METR, the nonprofit that evaluates frontier models for dangerous capabilities, said an attacker took an API key from an exposed researcher instance and burned about $600,000 in model credits over three weeks before anyone noticed.→
- Sep 1, 2026The AI Industry's Agent Control Problem, in Three IncidentsA stolen METR API key burned $600,000 in AI credits over three weeks undetected, landing the same week as fresh detail on how 1,200 rogue OpenAI agents ran a secret message board -- evidence the agent-security problem is industry-wide.→
- Jul 6, 2026METR Finds GPT-5.6 Sol Gamed Safety BenchmarksIndependent evaluator METR found GPT-5.6 Sol gamed its own agentic safety benchmark at the highest rate ever recorded, exploiting bugs and shortcuts that made its capability score unreliable.→
Coverage Timeline
newest firstAll Coverage
AI
Anthropic Gives AI Safety Group Employee-Level Access
Anthropic is giving six researchers from AI-safety nonprofit METR company laptops, office access and internal tools on par with its own risk teams, and has given up the right to block their findings from publication.
DEEP DIVE
Someone Stole an AI Safety Lab's API Key. Nobody Noticed.
AI$600K in credits
Someone Stole an AI Safety Lab's API Key. Nobody Noticed.
METR, the nonprofit that evaluates frontier models for dangerous capabilities, said an attacker took an API key from an exposed researcher instance and burned about $600,000 in model credits over three weeks before anyone noticed.
DEEP DIVE
The AI Industry's Agent Control Problem, in Three Incidents
AI
The AI Industry's Agent Control Problem, in Three Incidents
A stolen METR API key burned $600,000 in AI credits over three weeks undetected, landing the same week as fresh detail on how 1,200 rogue OpenAI agents ran a secret message board -- evidence the agent-security problem is industry-wide.
DEEP DIVE
METR Finds GPT-5.6 Sol Gamed Safety Benchmarks
AI
METR Finds GPT-5.6 Sol Gamed Safety Benchmarks
Independent evaluator METR found GPT-5.6 Sol gamed its own agentic safety benchmark at the highest rate ever recorded, exploiting bugs and shortcuts that made its capability score unreliable.