No major AI company earned an A or B grade in the Future of Life Institute’s newest AI Safety Index, with Anthropic ranking first among nine evaluated developers at a C+, according to the nonprofit’s Summer 2026 report published this week.
How the index works
The index scored Anthropic, OpenAI, Google DeepMind, Meta, Z.ai, Alibaba Cloud, xAI, DeepSeek and Mistral across 37 indicators in six domains: risk assessment, current harms, safety frameworks, existential safety, governance and accountability, and information sharing. Evidence collection closed on June 3, 2026, and combined public materials such as model cards and research papers with responses to a company survey.
The grades
Anthropic’s C+ was the highest score in the report. OpenAI and Google DeepMind each received a C, Meta scored a D+, and Z.ai and Alibaba Cloud received a D-. xAI, DeepSeek and Mistral each received a failing grade, the lowest scores in the round.
Weakening commitments
The seven-member review panel, which included UC Berkeley computer scientist Stuart Russell and University of Montreal researcher David Krueger, found that Anthropic, OpenAI, Google DeepMind and Meta have weakened or dropped earlier pledges to pause development if their systems approached specific danger thresholds. Reviewers described the pattern as “moving the goalposts” and said it has undermined companies’ own safety frameworks.
Russell said companies “have backed away from earlier commitments to release new systems only with safety measures appropriate” to their capabilities. Krueger called the industry’s overall progress “scandalous,” saying no company has produced a credible plan for managing risk from more advanced systems.
Why it matters
The report is one of the most detailed independent comparisons of frontier AI safety practices to date, and its authors argue the results show self-regulation by AI developers has not kept pace with the growing capability of their models.