Black Friday: Insider US$59 a year, regular US$79. Ends 29 Nov.→
Back Back
Models6/day by emailPrimary

Stanford's 2026 AI Index tracks how AI models perform on GPQA Diamond, a set of PhD-level chemistry and physics questions designed to resist Google searches. Mean model accuracy on the benchmark reached 93% in 2025, compared with an expert human validator baseline of 81.2% -- AI models now beat PhD-level experts by 12 points on GPQA Diamond

The Stanford HAI / Epoch AI 2026 AI Index tracks the performance of AI models on GPQA Diamond, a benchmark of PhD-level chemistry and physics questions. In 2025, the mean accuracy of AI models on this benchmark reached 93%. This compares to an expert human validator baseline of 81.2%, meaning AI models exceeded human performance by 12 percentage points.

SourceStanford HAI / Epoch AI·2026 AI Index

Verified 2026-10-01

Track with Insider

Why it matters

Models now handle expert-level science questions, but a test score is not the same as doing a scientist's job.

The AI boom,with receipts.

6 sourced facts every weekday, plus the Weekly Brief, AI company deep dives and so much more. Free.

Free · No card · Unsubscribe anytime · See today's issue →

Join now and you start with 100 Receipts on The Receipts Ledger — your first step toward a free year of Insider or Pro. How it works →