Skip to content
J·BJon BishopIndependent growth marketer

Research · AI-agent task-completion benchmark and results dataset

METR Time Horizons

agent capability progression, long-task automation analysis and model evaluation

Price
Free
Access
interactive chart plus YAML benchmark/result downloads

Put it to work

Potential use cases

A starting use for this source

agent capability progression, long-task automation analysis and model evaluation

Build a source-backed content brief

Identify relevant research or survey findings, then inspect the original source and population before using them in an article or marketing claim.

What it covers

Geography and granularity

Geography: frontier AI models evaluated on software/research tasks

Granularity: model, task, human task duration, success rate and fitted time horizon

Source type: AI-agent task-completion benchmark and results dataset

Access and cost

interactive chart plus YAML benchmark/result downloads

none for public data

Cost: free

Update cadence: new evaluation releases; downloadable versioned results

Before you use it

Limits that matter

suite coverage is limited, long horizons are unreliable and time horizon measures task difficulty rather than autonomous operating time

Prefer to write? Send Jon a message.