Back
Models6/day by emailReported
AI agents attempting OSWorld's real-world computer tasks across Ubuntu, Windows and macOS — file operations, multi-app workflows — historically topped out at just 1% to 12% success. Stanford's 2026 AI Index reports the best model, Claude Opus 4.5, now reaches 66.3% accuracy, within 6 percentage points of the 72.35% human baseline 66.3% accuracy on OSWorld's real-computer-task benchmark
Verified 2026-08-31
Track with InsiderTHE INTELLIGENCE CLUB
Facts like this, six days a week
The Daily Receipts on weekdays, The Weekly Brief on Saturday — six sourced facts a day by email, two more than the free site shows. Every one traced to the filing it came from. Free.
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 →
Read this morning's edition →The screener · 72 companiesEvery layer of the AI chain, priced dailyHow this was verifiedEvery figure links to the filing it came from
More on Models
- AI models can now win gold at the International Mathematical Olympiad, yet Stanford's 2026 AI Index found they still fumble a task most children master early. Across ClockBench's 180 clock designs and 720 questions, the top model, GPT-5.4 High, read analog clocks correctly only 50.6% of the time, versus 90.1% for humans
- 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
- Frontier AI models are catching up fast to benchmarks meant to resist them. A year ago they scored under 10% on Humanity's Last Exam, a 2,700-question test designed to be hard for AI and favorable to human experts. Now accuracy has reached 38.3% on Humanity's Last Exam, up from under 10% a year earlier