METR Q4 2026 Time Horizon - Higher Lifecycle Cost Tasks (50%): leaderboard

METR's Q4 2026 capability update (20 November 2026): the length of task, in hours of expert human work, that a model completes 50% of the time, fitted across tasks of different lengths. Fitted only on the half of the pooled TH1.1 + TH2 tasks that are costlier than their length predicts to build, run and score, by a model's cost estimate. Models score about 2.9 times lower here than on the cheaper half. Unbounded; higher is better. Reasoning effort and scaffold are not stated per model.

Metric: 50% Time Horizon (hours). Source: metr.org. Saturation forecast: Around December 2027. 15 models tracked. Updated 21 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-6 Astra23.97#10
2Claude Opus 5.520.79#9
3Claude Fable 5.113.16#20
4Claude Opus 512.83#26
5Claude Fable 511.35#33
6GPT-5.6 Sol9.53#37
7GPT-5.58.54#54
8Claude Opus 4.88.08#65
9Claude Opus 4.74.63#85
10GPT-5.43#135
11Claude Opus 4.62.83#113
12Claude Opus 4.52.2#160
13GPT-51.44#178
14O10.2064#403
15GPT-4o0.0613#751

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=metr-q4-2026-time-horizon-higher-lifecycle-cost-tasks-50 · How It Works · Data refreshed daily, snapshot 2026-11-21.