No AI Lab Scores Above C+ in FLI's Summer 2026 Safety Index
Summary: The Future of Life Institute evaluated nine major AI labs across six domains and 37 indicators — and awarded no A or B grades.
Key Facts
- Anthropic: C+ (1st); OpenAI and Google DeepMind: C (tied 2nd); Meta: D+
- Z.ai and Alibaba Cloud: D-; xAI, DeepSeek, and Mistral: F
- Graded by a seven-member panel of independent AI researchers using absolute performance standards, not curves
- Core finding: Anthropic, OpenAI, Google DeepMind, and Meta have all weakened or abandoned earlier pledges to halt development if systems approach specified danger thresholds
Why It Matters
The index is one of the first public, third-party audits of whether AI developers are keeping their own safety promises. The findings arrive weeks before the EU AI Act's high-risk obligations take effect on August 2, raising questions about whether voluntary commitments can serve as a credible governance baseline — or whether binding regulation is the only instrument with teeth.
Read More
- AI Safety Index — Summer 2026 — Future of Life Institute
- The Latest AI Safety Rankings Are In. Nobody Gets an A — Time