Nine frontier labs graded on safety; none scored above a C+
The Future of Life Institute's Summer 2026 AI Safety Index graded nine frontier companies across 37 indicators in six domains, with an independent panel of seven researchers and governance experts assigning domain grades. Anthropic topped the field at C+ (2.66/4.0). xAI, DeepSeek and Mistral failed outright.
The headline is that the best grade was a C+. The more useful detail is the distribution, because it shows the field is not uniformly weak — it is weak in a specific, structural way.
Anthropic led at 2.66 on a four-point scale, credited with comparatively strong transparency, an established safety framework, technical research and governance. OpenAI and Google DeepMind followed at 2.28 and 2.01. Meta took a D+. Z.ai and Alibaba Cloud sat at D−. xAI, DeepSeek and Mistral — one American, one Chinese, one European — received failing grades, which usefully spoils any story about this being a geography problem.
The methodology matters for how much weight to put on it: 37 indicators across six domains, absolute performance standards rather than a curve, discretionary weights applied by a panel of seven. Absolute standards are why nobody scored well. On a curve, somebody always looks fine.
A C+ from the safety-conscious lab is the finding. It says the leader is not close to the standard either, and the gap is not one company's failure to try.
Future of Life Institute — AI Safety Index — Summer 2026 → · Future of Life Institute — AI Safety Index Summer 2026 (PDF) → · TechTimes — AI Safety Grades Are In: No Lab Tops C+ →