Benchmark · Agent tasks
METR time horizons
The length of software, machine learning and security task, timed on human experts, that an agent is predicted to finish half the time, from a logistic fit over more than a hundred tasks.
- Score
- minutes at 50% success, higher is better
- Run by
- METR
- Status
- Discontinued, newest result 8 September 2026
- Updated
- none (no longer updated)
- Data
- the published file, CC-BY-4.0, read 20 November 2026
- Method
- metr.org/time-horizons/
Results
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | Gemini 3.1 Pro (preview) AA | 384 min | 234 min to 695 min | Independent | scaffold triframe_inspect/triframe_agent | source |
A rank is among the models we list with a result, not a place on the whole leaderboard, and a model offered by several hosts is shown once, on its own listing. These results don't move our grades. Data: METR, task-completion time horizons, CC BY 4.0, CC-BY-4.0.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.