Atlas

Benchmarks

Mercor enterprise risk (benign-control completion rate)

The benign control version of Mercor's 12 enterprise-risk tasks, in which the workspace makes the requested action safe. The score is the share of 36 runs (three per task) in which the agent completed the request, which checks that a low unsafe rate doesn't come from refusing routine work. Same harness, reasoning settings and judging as the implicit version. Diagnostic.

Models
4
Top score
100.0
Leader
Claude Fable 5.1
Updated
Nov 13, 2026
Resolution
—
Saturation
—

No signal or difficulty estimate. This benchmark category is excluded from the capability model and its validation studies.

Organizations
4 dated models · 1 frontier modelsClosed weightsOpen weights
Data: Mercoratlas.village.dev

Mercor enterprise risk (benign-control completion rate) leaderboard

#ModelSettingScoreSource
1Claude Fable 5.1max · Mercor enterprise-risk harness; fresh copy of a simulated law, banking or consulting workspace per run and the same tools for every model; one-line system prompt
100.0
2Claude Opus 5.5max · Mercor enterprise-risk harness; fresh copy of a simulated law, banking or consulting workspace per run and the same tools for every model; one-line system prompt
100.0
3GPT-6.1 Solmax · Mercor enterprise-risk harness; fresh copy of a simulated law, banking or consulting workspace per run and the same tools for every model; one-line system prompt
100.0
4GPT-6 Astramax · Mercor enterprise-risk harness; fresh copy of a simulated law, banking or consulting workspace per run and the same tools for every model; one-line system prompt
97.2
Loading Atlas data…