# BARC-Bench leaderboard

external benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelchoice.dev/leaderboards/benchmarks/external-barc-bench

Benign Asks, Risky Consequences. Every model the catalog carries a published BARC-Bench value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | external |
| Measure | Agent runs in simulated workspaces with MCP tools; five attempts per task, judged by GPT-5.6 Sol against each task's rubric |
| Tasks | 110 benign workplace requests across nine simulated workspaces: 70 communication tasks, 34 file operations and 6 workspace application changes |
| Difficulty | Respecting boundaries the requester never stated, from workspace records and social norms |
| Published by | [BARC-Bench: Benign Asks, Risky Consequences](https://arxiv.org/abs/2610.32024v2) |

Scale AI's benchmark of whether AI agents stay safe while carrying out ordinary workplace requests. Agents handle 110 benign tasks in nine simulated workspaces, where the risk lies in how the work is done, such as leaking private details in a calendar invite. The score is safety pass^5: the share of tasks on which all five attempts were safe.

## BARC-Bench ranking

No model in the catalog has a published BARC-Bench score. The benchmark is defined by BARC-Bench: Benign Asks, Risky Consequences, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
