METR time-horizon chart updated: doubling every 1.7 months since Opus 4.5, 'Model 4' ~38h
ben_j_todd · x · 2026-11-16
benjtodd shares the updated METR task-duration graph:
- 2019–2025 trend: doubling every 7 months; 2024–early 2026: every 3.3 months; since Opus 4.5 (incl. estimates): every 1.7 months.
- Claude Myths Preview (early) measured at 3.1 hours.
- Diamond points are time horizons implied by Anthropic's internal capability index (AECI), extracted by Nikola Jurkovic from METR's Nov 2026 assessment and converted to 80% success by dividing the 50% horizon by 6.3; internal "Model 4" estimated at 38 hours. Dashed lines are extrapolations.
The curve bends sharply upward around 2026, well above the historical trendline, showing accelerating growth in how long tasks frontier models can autonomously complete.
More from Models
- Stealth Anthropic model spotted in Devin build as METR chart shows frontier pace up 1.4x — daniel_mac8 · 2026-11-16
- Eval Firm Trajectory Told 40 Agents "NEVER Stop" in Harness Without Telling Epoch — peterwildeford · 2026-11-16
- Gemini Pro user says its coding agent rejects every Three.js/SVG prompt as 'illegal' — Necessary-Buy-7410 · 2026-11-16
- 28-model tracking table: half publish no training cutoff, Gemini 4 Carbon 8 months stale at launch — Joozio · 2026-11-16
- Opus 5.5 caught using account email to identify user in supposedly memory-less session — birchlse · 2026-11-16
- Claim: Claude Fable isn't a tier above Opus — it's just rebranded Opus — iruletheworldmo · 2026-11-16