Models will work autonomously for full 8-hour days by mid-2026
“Models will be able to autonomously work for full days (8 working hours) by mid-2026.”
Context: "Failing to understand the exponential, again" (Sep 27, 2025), extrapolating METR time-horizon and GDPval trends.
How we judge it
Our interpretation: by 30 Jun 2026, METR (or an equivalent evaluation) reports a frontier model with a 50% time horizon of at least 8 hours.
Resolution criteria are our interpretation of the claim, written before grading. Methodology
- Said
- Resolve by
- Verdict
- Correct ()
- Stated probability
- None stated
- Specificity
- 3 of 3 (precise and measurable)
- Score
- 1.0 points × weight 3
Evidence & grader’s note
METR's point estimate for Claude Opus 4.6 (14.5 hours) exceeds 8 hours, but the CI is wide (6-98 hrs) and METR calls the measurement extremely noisy. A 50% time horizon is not the same as working autonomously all day.
Evidence timeline
- Verdict Graded: Correct.
Dated entries from the claim’s own record (statement, verdict, later developments) and sourced entries added by editors. Each links to its source; none changes the verdict.
- Readiness milestone
- Part of AI agents complete multi-day tasks autonomously (a step toward it)
Share this scorecard
Post on X LinkedIn Share image (PNG)
Embed / Cite
The embedded card is live: it shows the current verdict, with the date the data was read. If a verdict is corrected later, the card changes too, so cite the page and its date for a fixed record.